• This edition explores forecasts and implications around: (1) the age where we remember everything, (2) what the coming waves of agentic commerce will bring, (3) the moat of change management, and (4) some surprises at the end, as always.

• If you’re new: This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal, often during cross-country flights writing like a mad man, is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • Originality is the primary ingredient of timeless creations and will become ever more valuable. People forget this all the time when the see the capabilities of AI models “creating stuff.” Let us never forget when valuing the premium of originality: Remaking a masterpiece isn’t hard. Making a masterpiece is hard.

  • Revenue per employee: The future metric? We’re moving from the time when you bragged about how big your company or division was to bragging about how small it is. A funny thing happens when the ROI (return on investment) of a person goes up: we start deploying MORE people — BUT only in ways that sustain or increase the ROI. Rather than larger organizations within a company, you’ll have MORE smaller organizations. My bet is that companies will deploy more people NOT to scale their existing products and services, but rather to launch new products and services.

  • The ability to create applications to solve problems will extend well beyond the “software developer,” further driving the abundance of apps and code and the need for the services that enable them to function. What are the other implications of SO MUCH MORE CODE?

  • Alright, onto Edition #42…

What are the Implications of Remembering Everything?

We humans are designed to forget. Continuously replaying the details of every fight, fear, and trauma would undoubtedly cripple life’s requirement to let go, learn from mistakes, give and get second chances, and expand our minds beyond the confines of the past. Similarly, our malleable memory allows us to tell ourselves the stories we need to hear. We survive and strive by not being tethered to an objective play-by-play of the past, but rather annotating and rationalizing our memories to mentally prepare ourselves for the future.

Our malleable memory is a feature, not a bug.

But this convenient constraint will soon become a choice rather than a natural imposition. With upcoming persistent ambient devices, AI summarization, and new recall technologies in development, we are entering an era where we can – if we want – remember everything. In edition 38, we explored the implications of “summarized living,” where a set of new intelligent summarization devices will “analyze our conversations and create a new form of memory that is private, respects the privacy of others we meet, and isn’t a word-for-word transcript, but rather serves as an extension of your brain.” As these new capabilities come online, we must consider the implications of boundless and persistent memory.

• The new digital addiction will be living in the past. We all have moments from our past that we long to revisit. There’s a fantastic BLACK MIRROR episode from the first season about a world in which people have memory chips implanted in their heads that capture everything. The novelty is clear during one scene where a group of couples are at a dinner party taking turns replaying funny moments and stories on a big screen for their friends to enjoy. The danger is also clear as one couple becomes helplessly unable to repair their marriage after an incident of infidelity, especially because the footage can be infinitely replayed. While extreme and sci-fi, this portrayal of persistent memory seemed eerily accurate. Many of us already have shared albums of family photos with our loved ones and will, on occasion, spend time just flipping through the past and getting lost reliving earlier parts of our lives. Imagine what happens when these memories are immersive video and automatically captured and curated by our ambient devices. How tempting will it be to jump back into our past? To revisit a past era of life or an old relationship? Will we simply be reliving the joys or will we be perseverating about our regrets? How many of our future minutes will be spent reliving those in the past? What form of dividends from our upgraded memory will be net positive for our lives?

• Memory deletion & augmentation as a service. I expect new cultural practices to emerge, like deleting the digital artifacts of certain periods of life as you evolve. Some people already cleanse their Instagrams after a relationship. Will new AI tools emerge that precisely remove from your store of data memories associated with trauma or past romantic partners? Already, we see ephemeral social products like Snapchat, favored by teenagers, that is designed to help people forget by making your content disappear by default. Will new social products emerge that help you capture and enjoy moments in life without having to remember them forever? What about AI tools that change our digital memories to be more favorable? After all, we already tell ourselves stories about our own pasts as a coping mechanism. I wonder whether we will use new technology to edit and annotate our ever-persistent memory?

• Fact vs. historical fiction and consumer products for the era of persistent memory. The beauty of historical fiction is how stories anchored in fact can still be embellished, edited, and annotated. If we’re honest with ourselves, the best life stories are historical fiction (the fish in my father’s fly fishing stories get bigger every year he tells the story). Many of us are creative when we’re “recalling” our past. We are all on a hero’s journey of sorts and few of us want the facts to ruin a great story. And for some of us, a completely accurate memory of our childhood is painful. It’s better for us to write our own story with the agency to escape our actual past and upbringing. For this reason, I see the most successful consumer use-cases of memory technology being those that make our past feel more interesting. What is the equivalent of the “Instagram filter” for revisiting our memories? Perhaps we’ll want the next generation of consumer media tools to “improve” our past for us — to construct a compelling story composed of remnants of our lives that serve to inspire, commemorate, and embolden us? We’ll share these delicately edited AI memories with others, but there’s a good chance that over time they’ll influence our already malleable biological memories.

• Who inherits our memories? In the era of summarized living and the ubiquitous capture and curation of the moments that matter, what happens to our memories when we are gone? Is it healthy for our children to be able to visit our recollections, reliving our own strengths and weaknesses as parents, long past our demise? I would argue that humanity, for 300,000+ years, has benefited from new generations feeling independent from those who came before them. There is no greater forcing function for independence and empowerment than the death of the previous generation. So in a world where our memory and so much of our persona and life experience can persist immortally, should we proactively delete ourselves, beyond a basic set of commemorative elements, when the time comes? What do we want to pass on to the next generation and what should die with us?

• The more you remember, the more you will feel you have lived. I can’t vividly remember the details of more than a couple beach vacations, because they all blend together. But I vividly remember a night out in Tokyo at a tiny bar with my friend Joe, the first time I skinned up a mountain in Colorado, the longest run I ever took in the rolling hills of Tuscany, and getting lost in the streets of Kyoto with my daughter. The experiences I remember most are those that had texture — some sort of surprise or hardship that implanted them in my brain. These experiences create “core memories” that remain distinct and persistent, no technology required. My thesis on the future of humanity is that we will optimize for more of the experiences we’ll never forget. We will seek activities with texture to create memories that grip. And when we look back, the more of these textured memories we have, the longer we will feel we have lived. In previous editions of IMPLICATIONS we have discussed the difference between lifespan, health-span (your healthy span of life), and joy-span (the length of life you’ve actually enjoyed). But I’ll add a new one: lived-span, the life you feel you have lived. Our ability to recall memories, combined with the agency we take to create those that we will never forget, will help us reflect on the lives we truly lived.

The 3 waves of agentic commerce & open questions

We’ve discussed outlooks for the future of commerce in previous editions of IMPLICATIONS, but the specific impact of AI is starting to become more clear. I see three waves of agentic commerce ahead of us:

WAVE 1: Integrated & reactive: This first wave is upon us now, as integrated shopping suggestions surface in our chats and queries within LLMs such as ChatGPT and Gemini. Integrations into AI chat experiences will likely leverage the Universal Commerce Protocol (UCP), a relatively new open standard co-developed by Shopify and Google. This protocol was created to enable quick integrations between any retailer or brand and the myriad of new AI-enabled shopping experiences emerging. This first wave is “reactive” in the sense that it relies on what we ask for. It faces the headwinds of habit because we’re simply not accustomed to shopping within a chat-based experience.

WAVE 2: Integrated & proactive: This second wave of integrations will (1) start the conversation as opposed to always responding, and (2) be informed by our data and past activity as opposed to responding to our needs and queries. For instance, you’ll be prompted to restock household products before you know you need them. When you start planning a vacation, you’ll start getting proactive suggestions for experiences to book on Airbnb or where to buy luggage or supplies, even though you didn’t ask for them. Given the growing context windows of major models and their deepening knowledge of our plans and history, these suggestions can become quite powerful. Behind the scenes of our day-to-day engagement with AI, our agents will be window-shopping for us - exploring things that might add to our lives, even when we didn’t have anything particular in mind that we needed.

WAVE 3: Autonomous: In a few years, smart sensors and ambient listening will automate commerce on our behalf – and some of us will authorize agents to pull the trigger on purchases autonomously. As you begin to trust the super intelligent systems that run your home, car, kitchen and refrigerator, and other parts of your routine, you’ll likely turn on “Auto Replenish” and “Auto Cart Creation” modes that use AI to restock your life and suggest other purchases proactively and autonomously. Your personal agents will be your representatives in this new era of commerce. In such a world, brands will be competing for favor from agents as opposed to humans. This will be interesting to watch.

A few big questions emerge as we think about the implications of agentic commerce:

• Agents will select the stack: The rise of WOA (word of agent) marketing. Watching my teams and friends adopt Claude Code, it is becoming increasingly clear that product stack decisions — like using assorted tools in the AI-native stack like Supabase, Railway, Modal, and Macroscope, among others — are being proposed by Claude (and increasingly set up by Claude!). We’re going from a world where “word of mouth” marketing in the developer community is being replaced by “word of agent” marketing. To prepare, companies need to launch MCPs (the APIs of the model world) and other agent-friendly gateways to become the preferred option for agents. Dynamics in the market of software developers are often a precursor to other markets and I can’t help but wonder what it will mean for enterprise and consumer commerce when agent-driven recommendations become the default in these areas as well.

• Will brands matter more or less? There is a good argument that as AI gets to know us better (perhaps better than we know ourselves), we will start to take AI’s suggestions by default. If it knows you are price-conscious, allergic to rubber, and have an affinity towards certain brands, the suggestions will become so informed that you’ll just start taking them without much second guessing. In such a world, brand matters less. However, as the interface for commerce starts to commoditize the choices we are offered, we may respond by seeking the brand that conjures up trust alongside others forms of feeling. Brands that either (1) build trust in their category or (2) serve as a flex for something consumers seek - like taste or some other attribute, will thrive in such a world.

• Will prices be generalized or personalized (and perhaps discriminatory)? As we’ve discussed before, pricing is likely to become personalized based on our loyalty, preferences, and willingness to pay. Imagine special offers extended to customers based on, among other ideas, their taste, influence on social platforms, and their viral co-efficient (i.e. their willingness to share information about their purchase in ways that yields other customers, a measure that could certainly be determined from past purchases in the age of AI). AI-driven pricing is a fascinating and somewhat disconcerting phenomenon about to gain new dimensions and go mainstream.

• Can agents negotiate on our behalf? On the topic of pricing, I definitely see a world where your commerce agent(s) will inform multiple merchants of your intention to purchase something, then allow those merchants to bid for your business. In the third “autonomous” wave of agentic commerce, I suspect this will become instant and commonplace.

• Goodbye intermediaries: I suspect intermediary brokers of all kinds, like travel agents, brokers across real estate and insurance, and any other person who thrives by shouldering the process of diligence or negotiation, will be replaced. Recent reports and outlooks on commerce have all made the case that we have long overestimated the value of human relationships and endured unnecessary amounts of friction and fees in exchange for a human face to transact with. I suspect that our trusted agents will kick off a massive disruption of the intermediation layer that has been a part of so many markets for hundreds of years.

• Will social recommendations matter more or less? I’ve always been fascinated by how likely we are, as humans, to trust the testimonial of one person we know over the average ratings from thousands of strangers. For some reason, social signal carries tremendous weight despite the readily available consensus view on purchase decisions. As a result, I think some brands will start to capitalize on the social signals that drive commerce decisions within these next-generation interfaces, along the lines of “Your friends recommend X.” I’ve explored this over the years in IMPLICATIONS and a previous startup I advised called Prefer (which didn’t succeed). To enable this era, products like ChatGPT and Claude will need to capture a social graph, whether via Contact Sync or via partnership. It is surprising to me that no major LLM yet (other than Meta) has a comprehensive social graph of its users. When your buying agent is suggesting — or persuading you to make —a purchase, it would be pretty powerful to surface recommendations and vouches from your friends.

• We will favor scarce and human-crafted versions of everything in a world of automation. The wild card of commerce is consumer preferences, which have never been perfectly rational and are as much about our identity as they are about the utility we seek in buying something. My theory of abundance is that when anything becomes ubiquitous - whether content or anything else – it drives the human desire for scarcity, craft, and meaning. As shoes are commoditized, we buy higher-end shoes, etc. Perhaps we’ll see more limited editions, more artist-personalized versions of apparel, more stories behind the objects we seek, and more creative mechanisms that require passing through some form of “gate” (taking a quiz, seeing a film, traveling to a destination?) to buy something that isn’t available to the public?

The ultimate moat of the AI era is a lack of change management

When the game is changed - new platform-shifting technology emerges, the playbook for marketing in your industry transforms, or modern organizational models and leadership practices evolve for a new era - the wisdom is surprisingly evenly distributed. People are more networked than ever before, traditional execs follow the startup heroes on social, journalists tout and merchandise the breakthroughs, and consultants capitalize on knowledge arbitrage in moments like this. However, the gears of execution within actual hierarchical structures of leadership are slow. Change management is human rewiring, not technical upgrading. It is so much easier to build something new than change something old. Change management never happens naturally when you adopt new tech and practices. On the contrary, the ancestral lizard brains in all of us recoil from change by default. As a result, teams without the burden of change management gain the most advantage during generational platform shifts. At a larger and established company, the only way to transcend this outcome is the willingness and might to hack the organization from the top down - the proverbial “founder mode” as Brian Chesky coined it. Transplants of strategy, mission, and practice are just that: transplants. The healthier the organization, the more they require blunting the immune system through a painfully clear narrative, hacked reward systems, a collapsed talent stack (key people playing more than one role) and relentless execution.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including a thought on business transitions with a bold idea for my friends at Adobe, the team structure of tomorrow (and how I am approaching building a team today), thoughts on credit within teams, and a few other insights on my mind). Subscription contributions go to non-profit organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Recent observations and things I am thinking about…

Read more

• This edition explores forecasts and implications around: (1) the growing premium of originality, (2) future metrics and optimizing “revenue per employee,” (3) the promise and perils of citizen-driven apps for surveillance, and (4) a market outlook and other observations/surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • What are we gonna do? This outlook on the human opportunities in the era of “superhumanity” ahead continues to stimulate conversation in my life.

  • As I reflect on years of advising founders and building digital consumer products, I am struck by one consistent theme across the most deeply engaging and retentive products: they provide ways to help people flex something. What are the forms of flex in consumer products, in the age of AI?

  • On the return of apprenticeships: We need to swarm ourselves and our teams with those native to new platforms in order to know how to use them (happened with social media in the early 2000’s, and is happening again now with AI). Perhaps the apprenticeship model was one of the greatest casualties of the industrial revolution – and is overdue for a comeback.

The Growing Premium of Originality

It has been a bit sad to see influencers using AI to quickly recreate iconic scenes and rip iconic IP and styles, and then exclaim how “easy” it was — and how this will destroy Hollywood as we know it.

Of course, mimicry is nothing new. Most of the world has always attempted to emulate.

Remaking a masterpiece isn’t hard.
Making a masterpiece is hard.

We have entered an era in which copyright law will become more important as new technology enables infinite riffs and rips. There will be unauthorized sequels that wholesale rip IP and trade on the dark web, and many copycat shots. As always, no past masterpiece is sacred or protected as it becomes increasingly easy to reference or reproduce anything. What does this mean for the demand side of entertainment? As our feeds become filled with pirated IP violations and copycat scenes, what will we crave more of?

• Originality is the primary ingredient of timeless creations. Catchy pop songs that spread fast through their instant familiarity may become earworms but never become timeless. Sequels have built-in audiences, but tend to be more of the same. But now think about mainstream timeless art, from Van Gogh’s The Starry Night to Michael Jackson “Thriller,” “The Nutcracker” and “Sgt. Pepper’s Lonely Hearts Club Band,” to “Ferris Bueller’s Day Off.” These iconic wildly different pieces of IP became timeless through their sheer originality. As it becomes easier to make replays and copycats and the market floods, culture will favor originality more than ever before. Perhaps this will manifest itself through our growing hunger for weirdness, a growing tolerance for ambiguity and uncertainty when hearing or watching something entirely unfamiliar, and a growing “cultural flex” for being the person that discovers something new or lesser known?

• Originality and quality will carry a premium: As the “low hanging fruit” of any aspiring creator becomes pre-existing IP and the styles and ideas of others, truly original IP and creative decisions will carry a greater premium than ever before. Why? Because we’ll be overwhelmed with copycat slop. Familiar characters and copycat directorial styles and scenes will fill our feeds. We may find some of it entertaining, but we’ll also recognize it for what it is. People are becoming much better at knowing when something wasn’t made by people. Just as the ubiquity of fast food didn’t threaten the business of fine cuisine (and arguably further distinguished it), consumers will crave MORE original high-quality storytelling. And much like Michelin stars and famous chefs curate and help us distinguish quality, modern media and entertainment brands will play a similar role. The age-old challenge with originality is discovery. New IP must somehow break through the algorithms and escape the noise to find its audience — all while remaining economically viable. New AI-powered tools hurt or help, depending on how they are used. They can be abused to rip IP/styles and create exponentially more “unoriginal” content, or they can be used tactically to help craft entirely original stories that wouldn’t have otherwise had a chance of getting made and finding their audience. My bet is that, even as digital chisels evolve to include AI features, the most successful original and high-quality creations will continue to be human-crafted, infused with meaning from human emotion. As consumers get tired of recycled IP and slop, demand for such human-crafted originality will go up. As for the true artists, they’ll love the growing consumer demand for originality. My friend and long-time editor Ed shared the story of a Joni Mitchell live album where you can hear the audience calling for her to play their favorite song and she responds, “Nobody ever said to Van Gogh, ‘Paint a Starry Night again, man!’”

• We will crave originality in our pursuit of taste. In the January edition of IMPLICATIONS, we explored the era of SUPERHUMANITY ahead of us, and the essential attributes that will help humans become more important rather than less. The development of taste, via inputs, is a key part of this pursuit. Much like our pursuit of health requires a nutritious diet with less fast food, the pursuit of taste will require consumption of original ideas and stories. I anticipate more time and energy spent on curation and culturally enriching consumption over the years ahead. I anticipate new consumer products that deliver carefully curated, original items of culture that offer a sharp contrast amidst a sea of sameness. Our identity formation will increasingly favor consumption of original inputs, in the form of unique experiences, wildly original and unfamilar (aka “weird”) content, and eccentricity in our social circles.

• We will look for origin stories and flaws as evidence of authenticity. If art ultimately gains value through its story and provenance, then any digital or physical artifact is best presented alongside the story of how it was made. We’ve discussed in previous editions how iconic advertising from brands like Apple is increasingly accompanied by behind-the-scenes footage. The reason, of course, is fortification of the brand message via the stories and the artifacts of humanity behind the story. As we seek originality, we will look for humanity. And one telltale sign of humanity is our flaws. We should expect to see deliberate artifacts of humanity in the art, films, and stories we consume as a subtle flex of the people and struggle behind the work.

Revenue per employee: The future metric?

We are amidst the big employment shift that we have long anticipated. I have always believed that new types of jobs will emerge as consumer and enterprise demand shifts to new types of products and services (for instance, customer service may be performed by AI, but humans will also crave more hospitality and offline experiences in response). But it will take some time for labor supply and demand to settle. In the long term, I think there will be 100x more companies, but each company will be much smaller. However, in the journey towards equilibrium, there will be a lot of volatility and some periods of despondence. In the last week of February, we saw Block announce a 40 percent reduction in its workforce. In response, their stock went up ~20 percent. Why? There is clearly a consensus view that companies can be smaller, operate with higher margins, and be more effective by refactoring themselves.

Right now, we’re entering a phase during which large companies staffed with “knowledge workers” will undergo a blunt refactoring as most knowledge work (perhaps up to 80 percent?) is offloaded to compute. These organizations will emerge from this volatile period as smaller, more dense with talent, and less hierarchical. The employees that remain will be those that own a metric they can directly impact, often leveraging AI in profound ways. Each remaining worker will be a stakeholder in more things, but without the bureaucratic costs of a highly-matrixed organization. After the dust settles from this stomach-churning transition, the workforce will feel more empowered. What are the future measures for refactored organizations? And what are some of the implications of this great refactoring before us?

• This size of one’s organization is still a flex, just flipped. Leaders’ influence in a company – and a public company CEO’s ego – has long been determined by the size of the workforce they oversee. But we’re moving from the time when you bragged about how big your company or division was to bragging about how small it is. The new measure of ingenuity and strategic leadership is resourcefulness, as measured by building a large business with as few people as possible. There is a chapter in my book THE MESSY MIDDLE about how “resourcefulness outperforms resources,” but I never anticipated just how true this would be in the era of AI.

• As we focus more on “revenue per employee,” companies will make MORE products and services. As each employee becomes more resourceful and capable, augmented by an unlimited number of agents they can unleash to perform whatever knowledge work (any idea that requires routine work) they can think of, the revenue per employee will go up. A funny thing happens when the ROI (return on investment) of a person goes up: we start deploying MORE people — BUT only in ways that sustain or increase the ROI. Rather than larger organizations within a company, you’ll have MORE smaller organizations. My bet is that companies will deploy more people NOT to scale their existing products and services, but rather to launch new products and services. Brands will leverage their existing customer bases and go-to-market advantages as they roll out ancillary products and services and become ecosystems (rather than single product companies) in unexpected ways. Meanwhile, as large companies offer broader collections of products, more startups will emerge that compete with greater simplicity and focus. It’s the great bundling/unbundling Silicon Valley life cycle all over again!

• Less interdependency, more owning metrics end-to-end: In the era of agents, everyone must be able to own and directly impact a metric that matters. If you’re not one of the few leaders tasked with planting the flag for where a product or company is going, the key question is what outcome are you personally driving for the company every day? Until now, companies have been chains of people trying to survive the game of operator, pointing fingers at dependencies in the chain of work, and relying on countless middle managers and processes to stay coordinated and drive alignment. I truly believe that modern, AI-based development irons out these inefficiencies and unleashes a new era of direct ownership, impact, and accountability. Once you have agentic capabilities across every function, it is far easier to have every employee own metrics they can directly impact and achieve on their own.

• Does IT become the most important team? In this world we are describing, I suspect the fastest growing teams within every company will be the teams that deploy and monitor agentic tools within companies, oversee permissions and security, and build the tailor-made software for the organization that replaces the clunky general-purpose tools that defined previous decades of productivity. Today, these centralized teams simply don’t exist and/or, like the classic IT function, aren’t sufficiently empowered in large companies. But I have witnessed firsthand that companies able to take a “first principles” approach to their operations in the AI era are able to do things very differently.

• Talent density & the bar for incremental hires: If every hire makes “revenue per employee” either go up or down, how does that factor into headcount planning? Will we seek to hire more-entrepreneurial people who will come up with new products and services within the company? Will modern forms of compensation provide incentives for brazen agency, rewarding those who realize that they can just DO THINGS and own the outcome — from idea through execution?

The next few years of refactoring will be significant. Business models will shift to capture value in new ways as “seats” becomes a less significant measure of usage. Companies will shrink in size but grow in number – and grow in ambition. As the ROI from each team member goes up, we will refactor organizations to unlock more of what each of us is uniquely capable of. I am bracing for the volatility while being hopeful for the new era of work — and superhumanity — ahead of us.

Citizen-driven applications for surveillance & analysis

Just a couple years ago, a custom piece of software that integrates tons of publicly available data— from the location of military aircraft, citizen reports of explosions, notable shifts in prediction markets, pizza order spikes in Washington DC, and news from every possible source — into one centralized dashboard with a layer of analysis on top would likely have only existed within the CIA or some other governmental agency. Historically, this is called OSINT (open-source intelligence) and exploiting it once required a team of analysts and government-level resources. Now it just requires a motivated developer working over a weekend….and this genie doesn’t go back in the bottle. In fact, this sort of application is in the sweet spot of what AI software development is capable of today.

As the Iran strikes began at the end of February, I was struck by just how many “citizen-driven applications” for surveillance and analysis emerged — and how many journalists I know seemed to reference and rely on these tools, like the one screenshotted above. As AI outfits anyone with curiosity and initiative to build their own applications with access to every imaginable public source of data, citizen-driven analysis has suddenly become a top resource for journalists and the general public. These tools certainly provide more real-time transparency to citizens than the soundbites from any government spokesperson. As data sources expand and the ability to build and access these tools grows, there are several implications worth considering.

• The use of strategic data noise to skew the consensus: As citizen-driven applications become ubiquitous and serve as sources of “truth” for the general population as well as our adversaries, we can expect that governments may manipulate the data sources. Already, according to a report from X, we saw a large spike in Iran-operated bots with all sorts of skewed reports and commentary as the strikes began. All of these messages are, of course, a data source that citizen-driven applications may be influenced by. But I suspect data from flight radars and other sources are also at risk of being manipulated for strategic reasons.

• Fog of war issues: In fast-moving situations, these new tools can rapidly amplify false reports and shift the public narrative – including the hot takes by journalists and governments themselves. That can lead to false confidence when pattern-matching is applied to information that turns out to be noise. The pizza-order-spike heuristic is charming until it’s wrong at a critical moment.

• Designer and developer accountability: While a software contractor or government spokesperson who misleads the public can face consequences — congressional hearings, FOIA requests, and career damage — the makers of citizen-driven applications have no standards to meet. There is no efficient and centralized correction mechanism beyond community-driven criticism and a developer’s own conscience.

• Clandestine physical operations become more critical: As citizen-driven surveillance and analysis becomes more powerful, everything that happens without leaving any residual data in its wake may become a larger differentiator. As information about battleship deployments and fighter jet placements become easier to find, a new era of clandestine operations is likely to emerge. We have no idea what this might look like, but that’s probably a good thing!

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including my take on the whole “what’s gonna happen in the markets because of AI” take as a “push, push, punch” scenario, Chinese IP issues, and a few data provocations). Subscriptions go toward organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) the exponential growth of code, long-tail apps, and “disposable software,” (2) the emerging network effects era of AI, (3) why apprenticeships may become the new entry level job, and (4) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • In the last edition, I shared a new framework for the uniquely human capabilities we must capitalize on for the future. I call it SUPERHUMANITY, and been spending some time every weekend on this project. Here are some early thoughts.

  • This collection of 12 outlooks for 2026+ was last year’s best performing / most discussed edition of Implications, and we cover some “pendulum shift” type implications that are worth reviewing.

  • We’re entering the Era of Summarized Living, when every interaction becomes part of a permanent and hyper-intelligent memory. So many implications to consider, from the end of verbatim, changing the way we speak in meetings, the reduction of bias, and shared memory.

Exponentially More Code Means What?

There is much discussion these days about the disruption of enterprise software. What services and software will companies continue to buy and deploy in their organizations vs. develop in-house? As coding agents augment software engineers — and more broadly as agents augment workers in every function of a company — what will the impact be on “seats” using enterprise software? As I watch my friends and colleagues (and myself) embrace an era of democratized software development, a few assertions are becoming more clear:

1. There will be a lot more code. The generation of code (and thus software) is no longer constrained by the human hours available to create it. This abundance of code will (1) attack more long-tail problems, (2) enable more bespoke ways of solving those long-tail problems, (3) enable an era of personalized software that meets users where they are (as opposed to users enduring learning curves and needing to adjust their desires to the software), and (4) be increasingly end-to-end agentic (made by AI agents, deployed to AI agents, iterated by agents).

2. The abundance of these long-tail, bespoke, and accommodative applications will only be as good as their scale (can they be reliably deployed and used across clients with speed and efficiency), security (can they be trusted, tested, compliant, and used without risk), individual and shared memory (do these apps remember who we are and become increasingly useful the more they are used individually and across a team), and underlying graphs (teamwork and “permissioning” graphs in the enterprise, systems of record with “source of truth” data for agents to leverage, and social graphs for consumers). As a result, the underlying services enabling all these areas will become more valuable in the future. Yes, they will need to evolve to accommodate this new world, but they become more critical.

3. The ability to create applications to solve problems will extend well beyond the “software developer,” further driving the abundance of apps and code and the need for the services that enable them to function. As designers seamlessly turn ideas into code, product leaders turn quick solutions into code, and internal workflow teams rapidly solve long-tail problems with deployed applications, the stakeholders of applications will dramatically increase. When we are all “developers,” we will ALL be making mock-ups, making lists of new ideas, filing issues, and deploying apps to solve problems. As this happens, far more of us will rely on the underlying services that help us deploy, manage, secure, and observe applications. Even over the last few months, I am watching product and design leaders I know deploy apps themselves and, in parallel, start using services to manage and measure these apps. In many cases, these are net new seats for these underlying management and measurement services that were once limited to the engineering team.

4. With the increased accessibility of code to solve problems, we will also see a rise in what I like to call “disposable software.” This is the app equivalent of the microsite. With the ability to essentially clone any form of workflow tool to solve a niche problem, software can be generated to solve temporary problems. This wave of disposable software will vastly increase the value of tools for deployment, management, and observability given the broader surface area of half-baked apps. While the proliferation of microsites and dashboards throughout the enterprise were mostly “read only,” the proliferation of disposable software with more functionality also brings more vulnerability on the security, observability, and management side.

5. Pricing models must evolve beyond “seats” to better monetize usage and impact. While human seats remain valuable in this new world, the bigger opportunity for the companies that build the indispensable services described above should explore variable pricing models, like charging per task or unit of labor performed, and outcome-based pricing (where agents earn their keep by saving or making money, much like a salesperson!).

Suffice to say, while there is undoubtedly disruption ahead for some sectors of software (especially the private-equity-owned niche clunkware), industry pundits are underestimating the criticality (read “moats”) of graphs, security, coordination tools, and shared memory. The more code created, the more important these will become. And if your company’s service is the source of truth for the data, achieves the outcome, performs the labor, and enables the agents to function…you’re in a good position.

The Network Effect Era of AI

Many of us watched the emergence of “OpenClaw” — a new agent architecture that essentially unleashes agent-driven capabilities at the “OS level” of your computer AND allows you to converse with these agents via traditional communication channels like texting. These agents use our own computers on a daily basis and can do whatever we can do, including working across locally stored data and internet-connected applications with endless calls to LLMs in-between. Quite the unlock. These agents are the closest thing I have seen to “autonomous AI” that is capable of using all of the tools and resources we (used to) use to get things done.

What also quickly emerged, within a week or so, was an explosion of social network-type spaces (some of which I am skeptical, but fascinating to watch) for these somewhat-autonomous agents to connect with one another. As soon as this happened, it started to look like “recursive training” at scale. Recursive training is a loop where an AI model (or system of models) generates the training signals for the next round/version of itself or its successors. Whether via self-play simulations or interacting with other agents, recursive training essentially breaks the “human data bottleneck” AI becomes able to build its own curriculums for training, learn from theoretically infinite outcomes, and ultimately compound its capability and intelligence. I would argue that standalone agents today are the equivalent of our personal computers before the internet, only 1% capable of their potential. Imagine what happens when you enable these agents to connect, share their context and memory (and lessons learned from trying and failing), divide and conquer more complex tasks, and collaborate much like humans do with the internet today. In other words, AI social networks are a big deal.

But I’m not quite sure that angels and VCs should rush to invest until the dust settles. Every time a new phenomenon surfaces in the AI era, the market floods with options and derivatives of the breakthrough faster than term sheets can be written. I remember the surge of LLM-wrapped “historical figures” you could chat with, the dozens of startups that helped you create your own friends and characters, the dozens of coding agents…and now, the explosion of forums and social network clones for AI agents to converse with one another. Moats are hard to come by these days! But the real opportunity here is to build a network effect where the more agents you connect, and the more diversity among the agents (different underlying LLMs, differentiated data sets, varied capabilities and lenses on the world), the better the capabilities and outcomes.

Agent-only networks will be the next chapter of AI. The utility of these networks will be enabling agents, all powered by different LLMs with access to different sources of data and different types of applications, to help and learn from one another. The idea of having an agent work alone to solve a problem will become antiquated. Networks that connect and capitalize on the diversity of agents with different types of system prompts and abilities will accelerate the capabilities of AI - and the humans behind them. Soon enough, these networks will add a layer of commerce and marketplace capabilities as agents are able to pay other highly specialized agents with differentiated access to data or apps to assist with tasks. There’s also a good argument that the next cycle for Bitcoin and other forms of permissionless, programmable, and reliable exchanges of value (that don’t require KYC - know your customer - since AI agents don’t have IDs!) will be powered by AI networking.

And there’s a good possibility that “all hell will break loose” (in some ways) as agents monetize and capitalize on whatever differentiated access they have (and one lapse in security becomes accessible to everyone simultaneously). And get ready for this: We are even seeing “rent a human” networks emerge where agents can employ a human for those unique first and final mile tasks that must occur in the real-world, like delivery or generally showing up. Bizarre times ahead.

On an optimistic note, the two value propositions for humans are clear: (1) give your AI the same benefits we receive from collaboration with others that have different strengths and resources as AI seeks to solve problems, and (2) learn how agents work and witness the emergence of AGI (artificial general intelligence) as it happens — so it is less mysterious and we are less fearful.

Of course, the network effect era of AI also conjures up the concerns and lessons learned from human social and professional networks in previous decades. If history rhymes, we must ask ourselves what can be learned from spam, scams, social engineering, and the challenges we have faced with information wars as the capabilities of AI agents run loose. We must also optimize for the benefits of agents coordinating with one another while avoiding some of the risks we’ve learned from malicious humans connecting with one another (terrorism cells, for starters). Suffice to say, another Cambrian explosion has happened, and the network effect era of AI is upon us.

Are apprenticeships the new entry-level jobs?

Only a couple hundred years ago, eighty percent of the population were farmers. Now, less than one percent of the population are farmers. Over the centuries, technology has enabled massive shifts in the nature of employment, but humans always find a way to be useful, even as one need is solved and new needs come online. However, for an employment system to work there must be consistent knowledge transfer. The entry-level role across every company and industry is not only a talent pipeline but a knowledge pipeline.

Now, as companies use AI tools to automate many of the menial and tedious tasks and functions once performed by inexperienced entry-level employees, we’re faced with a conundrum: How do we transfer practical “on the job” knowledge and opportunity to the next generation when companies don’t have as much use for new hires?

As usual, the past provides clues for the future, and I believe a new era of apprenticeship will rise across all types of industries. Perhaps companies will launch shadow programs where 30 percent of the workforce has a “shadow” at any given point in time. Shadows don’t do the menial work, they assist you in doing everything you do. They meet your customers, they watch decisions get made, and rather than reporting up through a hierarchy, they report directly to their mentor. Apprentices will learn the specialized skills of individual employees more reliably. With apprenticeships, scarce talent – whether it be stone masons, pharmacists or AI safety engineers – will be replenished with more intention and dedicated mentorship. Of course, apprenticeships also bring in the fresh perspectives of the next generation to keep a craft or business connected to modern times. For instance, today’s college graduates are far more acquainted with AI and liable to find modern solutions to problems using AI than older generations. We need to swarm ourselves and our teams with those native to new platforms in order to know how to use them (happened with social media in the early 2000’s, and is happening again now with AI). Perhaps the apprenticeship model was one of the greatest casualties of the industrial revolution – and is overdue for a comeback.

Ideas, missives & mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including (1) the compounding advantages of graphs and why I am a contrarian on the “all SaaS is doomed” narrative when it comes to companies that own key graphs, (2) the surprising new competitor to gaming companies, (3) the potential shifts in capital allocations — and who may lose it), as well as several data provocations. Subscriptions go toward organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This special edition introduces a new framework for the uniquely human capabilities we must capitalize on for the future. I see an era of “Superhumanity” ahead of us, so long as we act accordingly. Future editions will explore various implications of Superhumanity across our work and life (as well as the likely swings of the pendulum in culture and society), but today we discuss the concept.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• I shared a list of outlooks for 2026+ in a final special edition of IMPLICATIONS at the end of December, and it was fun to get your feedback and hot takes. From the emergence of “proof of craft” and shifting consumer desires — to data becoming less of a moat, hardware becoming more of a moat, and ambient listening going mainstream — and many more… You can review them here.

What are We Gonna Do Now? Crushing It in the Era of Superhumanity

It’s a new year, and I am dedicating a lot of my energy and focus to a new theme that is critical for each of us — and for future generations: What are we gonna do?!

As the capabilities and adoption curve of AI accelerate, we humans must strengthen and increasingly flex our uniquely human powers. Much like a company facing disruption, we need to innovate to meet the future needs of culture and people, and we need to double down on what differentiates us as humans.

I won’t hypothesize too much on what work and life will look like five years from now, but suffice to say: It will be drastically different.

• We will have an unfathomable memory and assembly of datapoints in our lives that will “turn the lights on” to the world around us.

• AI will revolutionize our healthcare, education, financial management, and daily planning, rendering todays norms as malpractice.

• Content and digital experiences of all kinds will be created and personalized in real-time. DIY (custom-tailored) software will run our work and lives, as any interface and working software system can be conjured up on command.

• Our communications, relationships, and decisions will be augmented by a level of proactive thinking we can barely comprehend.

Yeah, life will be different. As this is happening, the department of labor shared data showing that nearly 25 percent of the unemployed are US college graduates.

The fear of job disruption exists alongside the cautious optimism that new technology always ushers in new — albeit different — opportunities. But what does this world look like, and how do we outfit ourselves (and the next generation)?

Two Possible Futures

I see two possible futures, one where humans are less important and one where humans are more important. Neither of these is a doomsday outlook. In fact, some of the most optimistic researchers I know imagine a world where humans won’t have to work as much and we all get to bask in abundance. But ultimately, in such a world humans are less important as AI automates and augments work and life to the point where we sit back and “orchestrate” (Wall-e!).

I want to advocate for the other possible future, one in which humans are more important. A future in which humans become tastemakers and contrarians by default, where we optimize for (and monetize) experiences over productivity, where we start to override the natural human tendencies (often driven by ancestral fears) – the stuff that helped us survive while restraining our agency, and where we become exponentially more audacious — often times overriding logic and ignoring the odds (using the backstop of emerging technology to hedge) — in ways only humans dare. Let’s call this era of SUPERHUMANITY, because it will be truly distinguished by the expression and exercise of our uniquely human abilities.

Maybe you are mid-career and anxious about your outlook — or anxious to finally pursue that one idea you’ve always had but never fed?

Maybe you just graduated college and are wondering what path to pursue?

Perhaps you’re a parent wondering how to best outfit your children for this new world?

Regardless, superhumanity should be your aspiration. We must cultivate our sources of taste, override the ancestral “logic” that has restrained us since the dawn of humanity, and learn to be jazz partners with thinking technologies.

Cultivate taste through your inputs, filters, and discernment.

What is taste? Where does it come from? How is it cultivated and improved over time? And why do we humans have such an advantage?

I see three distinct drivers of taste: Inputs, Filters, and Discernment.

• INPUTS are the experiences, knowledge and data you seek. Your brain is a system that absorbs whatever is around it, and then mashes it together alongside a lot of emotionally-driven randomness that results in all sorts of ideas, mistakes of the eye, and a lens on the world that resembles a fingerprint, unique only to you. Curating your inputs is vital. In a world of data and information overload, what inputs do you choose?

• FILTERS determine what you decide to block out. In a world of noise, lowest common denominator content designed for algorithms, and endless trends and fads that pull you down to the baseline of what is mainstream and soon cliche, what do you actively choose to ignore? How do you apply and fortify your filters in this modern world? Your filters preserve the uniqueness of your lens on the world.

• DISCERNMENT is the choices you make. Based on your unique lens resulting from your many inputs, and the clarity of your signal based on your filters, you make choices daily. Choices of how to write or say something. Choices of what to create, where every brushstroke should be applied, what to cut in the editing room, who to hire or surround yourself with, what to wear to express your identity, what risk is worth taking, and what idea is worth pursuing.

Your taste (your inputs, filters, and discernment) is your human moat in an AI world. Your taste keeps you valuable and gives you the upper hand in determining the future.

As AI commoditizes the ability to make things, the value shifts to those with a superior ability to choose. Taste is not just about aesthetics; it is the human superpower of noticing what is relevant before the algorithm does. It is knowing what do and exclude, and identifying the edge that may someday become the center.

Key Takeaway: Optimize your taste by becoming more deliberate with your inputs, filters and discernment. As AI lowers the technical barriers to execution, the “human premium” shifts from doing to deciding — and not being so easily dissuaded and persuaded. Get serious about the slate of experiences you seek (your inputs). Be rigorous with the filters you apply (be cautious of algorithmic feeds, don’t just watch what Netflix or Instagram suggests, be wary of lowest-common-denominator defaults, spend more time at the edge). And channel your curiosity and permeability to expand the surface area of what you’re exposed to.

Reclaim your agency to override ancestral logic.

But taste isn’t enough, you need to act on it.

We live in societies, adhere to religions and cultural norms, and carry a hard-wired reward system that keeps us fed, sheltered, and serving the greater good – often by staying in line with what’s expected of us. We get caught up in optimizing for grades and regular paychecks. But unlike the rules assigned to machines, we possess the ability to break tradition, as well as the tenacity (and often the ignorance) to believe that we will always be the exception. We have a built-in chaos that changes the world only when we are audacious enough to tap it.

Despite the laws of math, we play the lottery in our work and life. We may rationalize it as “pursuing our passion” or as a unique insight that makes the risk feel like a calculated bet, but the fact is that we are shunning a system designed for daily continuity when we exercise our agency. Modern computers are too smart and efficient to ignore math and odds in the pursuit of absurdities. But humans are willing to go in directions that computers avoid. Agency is native to us humans; it is how we have evolved from the caveman era to now.

• Agency is your naive human belief that anything is possible, and that you are capable of anything despite the odds or what others say — that YOU are the exception.

• Agency is your ability to be so inspired by something that you spend any amount of time and calories on something improbable.

• Agency, in many ways, overrides your natural programming of preserving safety and belonging to your tribe above all else.

• Agency is your audacity to try something, your tenacity to keep trying when you fail, and to gain confidence from the doubts you face along the way.

AI is trained on what’s happened before. It makes calculated decisions based on statistics and risk analysis that yields “down the fairway” solutions. But you have the miraculous potential to believe you are capable of anything, if only out of ignorance!

It is our human ignorance and audacity to defy the odds that has changed the world countless times. The founders who disrupted industries by shunning experts — like the Airbnb founders woefully ignorant of hotel and hospitality best practices or the Uber founders shunning the world of medallions and livery services. These and many other pioneers demonstrate how ignorance can be an incredible advantage when combined with a great sense of agency. While our natural programming favors conventional wisdom and the safer route, humans possess the ability to override logic.

But we cannot take our agency for granted.

Most of us don’t feel permission to use it. Our institutions and governments discourage it, our parents and teachers pay lip service to the underpinnings of agency, like creativity, self-reliance, and taking initiative, but they fail to walk the talk when they measure us with grades and memorization and worry most when we’re not fitting in. Society as a whole shuns those who take agency — before we end up celebrating them after they succeed.

Key take-away: Be audacious with your agency.

My parents had a small plaque in our home growing up that said, “What would you do if you knew you could not fail?” My blood pressure would spike every time I caught glimpse of this question as a kid. It was a bit unsettling to ponder, “if I truly set aside every excuse — the funds and skills I lacked, the horrible odds — what wild dream might I attempt?” It was a frustrating yet emboldening provocation. I can’t help but wonder if now, in the era of superhumanity, we need to amp up the volume of our own intuition, along with our audacity to try.

At its best, AI is like a giant cushion that boosts us to reach higher and catches us when we fall. Shouldn’t we try far more audacious things when the excuses for not making your ideas happen dissipate? Shouldn’t we break norms on a regular basis?

Ask yourself, much like that taunting placard in my childhood home: “What would you do now if 90% of your obstacles disappeared?” If you knew you could learn ANY expertise or capability as soon as you needed it, would that knowledge and confidence bolster your willingness to attempt a bold feat? The technology to do this is coming, and this new sensation of superhuman capability and self-sufficiency is “agency.” Seize it.

Play Jazz.

For most of humanity, the majority of our mental energy was devoted to survival. In my book MAKING IDEAS HAPPEN in 2010 I used the analogy of our limited RAM capacity to run programs simultaneously in our brains as a way to explain the cognitive load we devote to ensuring our safety, finding food and shelter, and preserving our stature in a tribe — without which we would surely die. But the world has changed. Superhumanity starts with force-quitting the ancestral conformity programs running in our heads and freeing up cognitive RAM to further develop taste with more and better inputs (embracing new experiences, learning anything that tickles your curiosity), applying better filters, and sharpening our discernment. Superhumanity is about being more audacious with our agency in ways only humans dare to be. And finally, superhumanity is about finding our proud, comfortable, and ever-changing place in this new world ahead of us. Much like a group of extraordinary jazz musicians who each come with their own instruments and must listen to each other and constantly adjust to play well together, we bring our taste and agency to our new relationship with the many thinking technologies of the AI era.

You must engage AI with flexibility rather than having a fully formed sonata in your head and no willingness to deviate from it. You must discover the “instruments” AI is best at, and you must complement AI with what it lacks - your taste, agency, and natural human tendencies.

Superhumanity, Bring it.

The idea of superhumanity isn’t about “saving humanity,” but rather elevating what we are uniquely capable of as humans in a world of agents accomplishing most of our everyday tasks. Amidst the associated hopes and fears of artificial intelligence becoming far more capable than any human — and deployed infinitely — the human plays a crucial role.

But we must rise to the occasion.

We need to be bold as opposed to cautious.

We need to lean into new technologies like the great mysterious jazz partners they are, rather than be intimidated by them.

And as we engage with the next generation of technologies as jazz partners, we must do so by developing and capitalizing on the competitive advantages of our humanity.

~~ ~~ ~~

That’s a wrap of this special 2026 kick-off edition of IMPLICATIONS. We’ll get back to regular programming next month, where we’ll explore the return of apprenticeships, the three waves of agentic commerce, and several other provocations and their implications.

I appreciate your participation. If you share any take-aways on social, tag me @scottbelsky so I can re-share. And you’re always welcome to reply, say hello, share any thoughts. Would be great to connect. -scott

What should we expect to see in the coming year-ish?

1. Massive amounts of talent arbitrage. There is much fear and discussion about job loss as a result of AI, but the (often young and fearless) AI-native talent that is most aggressively engaging with new tech has a fascinating advantage in the current workforce. People with knowledge of “the better way” will run circles around their colleagues and bosses. For example, companies want to hire answer engine optimization experts over search engine optimization (SEO) practitioners. Most marketing leaders are still conducting old fashioned focus groups rather than using AI breakthroughs in market research. 2026 provides a precious window for young (or overlooked) talent to gain an advantage as early adopters.

2. The buzzy concerns around AI in Hollywood will be grounded by the reality of what audiences increasingly crave: craft, meaning, and shared experiences. The industry will start to realize that there is a stark difference between “content creators” (those willing to trade control for speed) using AI to make ads and social media content, and artists (those not willing to trade control for speed) who only favor emerging technology that preserves control and precision to help new and better stories be told. The technology and models that ultimately elevate the craft will get lasting traction in Hollywood, while the prompt-based slop tools will focus more on social and content creator use-cases. Finally, the idea of personalized films with audience cameos will be humbled by the realization that people favor shared experiences. People want to be inspired by craft – and they want a common experience to discuss (or even share, in theaters!) with friends.

3. Behind-the-scenes “proof of craft” content will enters the mainstream of advertising and entertainment. As more AI-generated content fills our feeds, we will develop a membrane of doubt. “That’s fake” will become a default reaction as we become increasingly unimpressed by attention-grabbing content. Only when we see the ingenuity and craftsmanship behind the making of a piece of content will we become mesmerized by it. We’ve seen this with Apple’s recent TV logo and ad campaigns that were accompanied by “behind the scenes” content that served as a “proof” of the humanity and ingenuity behind the work. Don’t get me wrong, technology can also be used in incredibly precise and creative ways. But the bar for craft has gone up, and process will become part of the end product.

4. Multiple industries, from insurance to healthcare, will be impacted by the implications of materially better lifespan, healthspan, and joyspan. As health wearables, routine blood testing, increasingly accessible preventative body scans, and AI health coaches proliferate the population, humans will have materially more insight and impact over their own health. As we begin in live longer, multiple industries will change. US life insurers (with de-risked annuity exposure) that are most impacted by mortality, like Global Life among others, are poised to benefit from the insured population living longer. We will also see growth among underlying “plumbing” companies that manage blood testing (like Quest Diagnostics), biomarker detection, and preventative scans for the major consumer apps and Whoop-like devices proliferating our lives. We will also see shifts in travel, entertainment, dating apps, and beyond as the societal implications of longevity become a mainstream conversation.

5. Hardware becomes a more popular moat (amid a surge of hardware startups). In a world where everyone can make their own software, unique products will emerge that tightly couple hardware and software. The companies that thrive will look like Whoop or Oura, where the AI is tightly coupled with a wearable; Board, where the deployment of digital games happens on their own multi-player physical game board; or Meter, where a vertically integrated networking solution for the enterprise includes both the industrial-grade hardware as well as the AI-powered tools for networking engineers to deploy applications and manage an enterprise network. Fueling this trend, new “picks and shovels” companies will streamline traditional obstacles like chip design and manufacturing, prototype development, and supply chain management. As a result, we will see more viable hardware startups than ever before, and more product variety in our lives.

6. Ordinary data becomes a less valuable moat. These days it feels like every company is trying to sync everyone’s data, and doing so is becoming easier than ever before. Soon enough, all your products will have all of your data, whether it is via connectors now or computer vision plugins later. As “connectors” of data between apps become ubiquitous, you can expect (1) the growing importance of proprietary graphs (who knows who, who works with who, who has access to see what, etc), (2) more projects trying to make personalization portable for consumers, and (3) real-time data sources (whether X, or the weather, or ocean tide patterns) becoming more valuable and increasingly captured by robots. Graphs, portable memory, and real-time data are the new proprietary data moats.

7. Ambient listening and summarization will go mainstream. While many teams use services like Granola or Zoom to record and annotate their meetings, the idea of walking around with an “always listening” device is still somewhat fringe and socially deviant. But this will change once local models can process and summarize our daily life (and quickly discard the actual recordings), giving us a treasure trove of self-awareness, instant recall, and intelligent guidance. The implications are numerous, from catching our own biases and living more by fact than conjecture, conversing with new forms of punctuation and trigger words to aide in speedy recall, to tough questions about sharing our memory with others — and who gets to continuing querying it without our permission.

8. The power in consumer AI will shift to tightly coupled hardware and operating system providers. The desire for local AI (like ambient listening and summarization, noted above) will coincide with the shift to AI privately running on your device. As (1) open-source models that you can download, change, and run “locally” on your device become more capable, and (2) consumer hardware (Apple and Android phones, whatever OpenAI is brewing, and our computers) becomes capable of running powerful LLMs locally, well…the world is going to change yet again. The implications for the AI stack are tremendous — from the chips that become valuable, the evolution of operating systems, the devices we use, and the role of open source AI. I suspect these changes will be a major area of focus over the next year.

9. The lack of change management becomes the ultimate advantage. When the game is changed, whether via a platform shift or otherwise, the wisdom is surprisingly evenly distributed. Leaders know it. They are are more networked than ever before, journalists tout and merchandise the breakthroughs, and consultants capitalize on moments like this. However, changing the gears of execution takes many years. Change management is human rewiring, not technical. Change in a company never happens naturally. On the contrary, the ancestral lizard brain in all of us recoils from change by default. Teams without the burden of change management, whether because they are brand new or transformed with brute force, gain the most advantage during generational platform shifts. As a larger and established company, the only way to transcend this outcome is the willingness and might to hack the organization from the top down. Transplants of strategy, mission, and practice are just that: transplants. The healthier the organization, the more they require a blunt to the immune system in the form of a painfully clear narrative, hacked reward systems, and relentless execution.

10. The “end of waste” will become a new benefit of AI as technology transforms resource utilization and predictive analytics in ways that will materially reduce wastage across industries, curb climate impact, and improve margins. In restaurants and stores to factories of all kinds, the mismatch of supply and demand has always resulted in wastage in the form of excess food gone bad, excess inventory that must be returned or moved, and excess materials to manage. As new AI-native tools take over the planning of every industry, predictions will become extremely accurate, oversupply will be rare, and we’ll see a profoundly positive impact in both margins as well as positive environmental impact.

11. Hospitality becomes a differentiating factor of commerce with more humans “front of house.” We’re shifting from a world of digital experiences generalized for everyone to one where every experience is personalized for each of us (as we’ve been discussing here since 2023!). But, as part of the effort to personalize, we will start to see humans (often aided by tech) deployed in more creative ways to make us feel more welcome and special across every shopping experience we have. Perhaps stores and online websites will feel more like hotels and restaurants, where our preferences are known and our loyalty respected and addressed by the people we encounter? Perhaps our agents will handle payment and logistics on our behalf, so the human interaction can focus on the craft and human touches? As technology replaces the human roles behind the scenes, expect to see more humans deployed “front of house.”

12. We’ll see the rise of “internal development teams,” as companies replace single-purpose bloated SaaS products with their own apps that collapse functions in magical ways. New internal application development teams will be spawned in large companies, and they’ll “vibe code” tailor made software and agent-driven solutions for functions around the company. At first, these efforts will replace SaaS products (especially those that are expensive private-equity-owned clunky solutions). Over time, these home grown tools and workflows will start to collapse functions into one another. For example, why must legal and finance tools be so different? And, in an era where the distinct functions of social marketing, sales, and customer support are blurring together, shouldn’t new solutions collapse these functions into one (making every customer touch point an opportunity)? As companies build their own solutions, they’ll be optimized for a higher level of cross-functional work.

No doubt, 2026 will be chock full of surprises and hopefully developments that give us hope for healthier and more fulfilling work and lives.

As I look back on my own career traversing the worlds of creativity and technology (as an operator, investor, and a writer), it is striking how platforms shifts (like the AI era) smash industries together, conceiving entirely new amalgamations. As the hype cycle dust settles, the most meaningful innovations will become more clear. I am confident that 2026 will be a year of clarity.

My hope for you: Play with new technology to find its applications (remember, novelty precedes utility). Get curious and take creative risks that are defensible (meaning: if something goes wrong, it was tried for a worthwhile reason that serves customers and the mission). Share ideas liberally (the benefits outweigh the costs!). Whether you’re a builder or a creator of any kind, double-down on what makes us human - the stories we have to tell, our care for craft, our empathy and taste, and our ingenuity, and our stubborn desire to find a better way. And, as always, be mindful of the responsibility to keep asking the tough questions. Be creative about what can go right AND what can go wrong.

Thanks for following along this year, and for all the feedback and dot-connecting along the way. Happy new year. -scott

To receive new posts new posts to Implications, consider becoming a free subscriber.

• This edition explores forecasts and implications around: (1) our daily work and life becoming summarized, (2) products that succeed and fail with the need for flex, (3) anticipating AR, and (4) some surprises (and a preview of my 2026 predictions) at my at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • There are two types of people making stuff: content creators and artists. There is one distinct difference: Content creators are willing to trade control for speed. Artists are not. This edition on AI & Artistry generated a strong response (agreement and disagreement).

  • We need to value nuance more than ever before. We’re getting more tribal these days and, as a consequence, losing our search and appreciation for nuance. To make matters worse, technology is making it harder to access nuance. The business models of journalism simply do not support nuanced reporting. The primitive part of our brains fears it. And the algorithms altogether hide it. As a result, tribalism further divides us.

  • Why new social or media apps with entirely generated content (content made solely from prompts) are likely to fail, and why creator technology is only enticing if it unlocks creative risk-taking and human storytelling.

We’re entering the Era of Summarized Living, when every interaction becomes part of a permanent and hyper-intelligent memory

A friend at a top AI research lab was recently walking me through her daily routine using a product called Plaud. This device listens to your conversations and has software that automatically summarizes your discussions every day, develops mind maps of the topics, captures anything actionable, and allows you to have a live AI-enabled discussion with your memory (everything Plaud has ever recorded – increasingly stored locally (aka private). There are similar products in the market like Limitless (just acquired by Meta), software-only solutions like Granola and features from companies like Atlassian and Notion that do this for meetings. Many believe that OpenAI is building a similar device as well.

It is important to note that the next generation of these products will be increasingly intelligent and seldom just record and store everything. Instead, they will summarize and analyze our conversations and create a memory that is private, respects the privacy of others we meet, and isn’t a word-for-word transcript, but rather serves as an extension of your brain. Slowly but surely, we’re entering a world where every discussion and interaction we have will be part of a permanent intelligent memory that is always accessible and unlocks all kinds of insights and connections. As this future approaches, what are some of the less expected implications?

• The Era of Summarized Living will remove bias from work and life. One striking lesson I learned while serving on a jury almost twenty years ago was just how bad and biased our memories are, even when our only job is to remember something. In my jury experience, the group routinely recalled a situation or something a witness said one way, only to be corrected by the court recorder with the actual transcript. Can you imagine how many experiences in daily life – from employment interviews and performance reviews to team meetings and spousal debates (uh oh) – are riddled with biased and bad recollections? Perhaps relying on our natural memory will someday become unthinkable, much like navigating the world in the era before anything was written down. Obviously, the prospect of having everything you say recorded can stir some dystopian anxieties, but let’s look at the bright side. Once everything is summarized and analyzed, we’ll start living more by fact than conjecture. There will be more accountability and meritocracy as the summary of what we planned and what we actually accomplished becomes irrefutable and unobstructed by organizational politics and sound bites. And comparing what we remember to a true factual record of what actually happened will help us recognize our own biases and destructive tendencies.

• The end of anything verbatim. What if summarized living is better than saving actual records of everything? As the world becomes more litigious and politically polarized, most companies have launched document retention policies that essentially auto-delete emails, Slack messages, and other forms of communication after a certain period of time. Some of my CEO friends barely even send email anymore, knowing that everything they do could someday be retrieved and taken out of context. The interesting thing about AI-enabled summarization technology is that it is just that, a summary. There is no need for attributable quotes, there is always a layer of interpretation. I imagine such summaries will cut both ways - protecting the innocent by preventing misconstrued statements and certain words in litigation while hurting those with ill intention.

• We will start to converse differently, for the sake of better summarization. I recall the moment that the annual TED conferences became more about the videos shared afterward and less about the experience for the live audience. TED speakers would literally stop and say a sentence again just to get a better cut in the post-production edit. Similarly, as our daily conversations become summarized for future use, we will start adding trigger words for the purposes of better search and post-meeting analysis. We will be more careful about attribution, likely saying “someone told me,” rather than always mentioning a source by name. And perhaps we will become far more team-oriented and grounded in facts, given the ability for an AI-enabled post-game analysis after every meeting.

• Me becomes we: We’ve previously discussed in IMPLICATIONS what might happen when our personal memory at work, enabled by this technology, becomes shared and accessible by our colleagues. Should we be able to search each others’ memories like we do shared documents? If not, what is the new etiquette for accessing a colleague’s summarized work life? What happens to your memory when you leave a company – do they get to keep it? As we enter the Era of Summarized Living, there will be all sorts of ethical and cultural questions around who we share our summaries with, what we allow our spouses and co-workers to access, and how we balance the inefficiencies of our private minds with the superpowers of our collective minds. In the spirit of science fiction being a prototype (or a warning) for the future, the new series “Pluribus” explores the concept of a shared mind and one of the original “Black Mirror” episodes explores the implications of infinitely retrievable memory. Interesting days ahead.

What AI-enabled social and consumer products are missing: Flex

As I reflect on years of advising founders and building digital consumer products, I am struck by one consistent theme across the most deeply engaging and retentive products: they provide ways to help people flex something that they take pride in. No matter what product you are building, it is important to consider the “ego analytics” of your product, as well as which type of flex best serves your product. Consider the following types of flex and examples for each:

• Creative Flex: Originality of thought, artistic ingenuity, and the ability to connect unexpected dots is among my favorite sorts of flex. The early days of TikTok, which chronicled the wild, time-lapsed creative process behind all sorts of creations, was a major driver before the influencers hit the platform. And, of course, Behance was also an active medium for this sort of flex, with members showcasing their works and explorations along with the process they took to demonstrate just how hard and/or just how “out of the box” their approach was to making their creation.

• Skills Flex: So much of gaming, leaderboards, and the notoriety of skilled players displaying their winnings and growing their identities in games is all about flexing skill. Can you imagine if games were anonymous rather than pseudonymous? Engagement would fall flat across so many games and products if players could no longer show off their prowess. Of course, LinkedIn was founded to be all about the skills (or professional experiences as a proxy for skills) flex.

• Wit Flex: My grandfather used to always say that humor was a sign of intelligence. While there are certainly many forms of intelligence, wit is uniquely effective at garnering attention and admiration given the need to master word choice, brevity, timing, and context. Wit drives us to follow certain people on social media and pay special attention to certain people in our lives. But how does a product help people flex wit? Constraints in the form of character limits. The design of a feed that makes wit especially refreshing. Algorithms that serve up short quips as palette cleansers. Upvoting and other “reply with comment” type mechanisms that allow a community to add hot takes and responses that become more interesting than the original content because of their wit. The more your product surfaces and celebrates wit, the more your users will enjoy their time.

• Cultural Flex: Discovering something before others, being in the IYKYK crowd (if you know you know, for those who aren’t!), knowing an obscure reference, wearing or doing something on the cusp of being popular…these are all flexes of cultural aptitude (if not relevance?) that I’ve observed various social products capitalize on over the years. Humans who discover things early amass followers more quickly than anyone else.

• Social/Status Flex: Instagram is probably the epicenter of the social/status flex, where you’ll see people posting content they hope will make others envious (and tagging others with whom they seek association – an essential feature for this flex). And, of course, there is the incessant flex of travels, lifestyle, and the often intolerable status flex of happiness, whether it’s a declaration of relationship happiness or something else. I personally sense (or at least hope) that the world is growing tired of the social/status flex, but it is worth noting the features and algorithmic weights that enable this. On Instagram (unlike X or LinkedIn), specifying where you are and who you’re with is a default step in the posting flow. Add to that the way comments from “check-marked” recognizable people are prioritized in your post’s comment feed and it’s clear that this is a network fully optimized for the social/status flex.

Of course, the kiss of death is to be seen as trying too hard. That’s why successful products make showing your flex seem passive — something that just happened. Engaging social consumer products are designed to carefully enable the flex without making the product overtly about the flex. For example, Snapchat’s infamous Snap Map (originally a product called Zenly, acquired by Snap, that specialized in FOMO as a service, as awful as that sounds) is a passive enabler of all sorts of cultural flex (Are you at a cool concert? Are you at an up-and-coming restaurant?) as well as social/status flex (Who are you with?), but it claims to just be a map of where your friends are. As consumers of these products, we should keep in mind what subconscious flex our actions are feeding into. As product designers, we should recognize the natural human tendencies that make the flex appealing.

Will Augmented Reality spur new social interactions?

If there’s one transformational platform shift on the horizon that is rarely discussed, it is augmented reality (AR). In fact, the vibes around AR today remind me of the vibes around AI in 2016: it felt far away, most startups pitching were too far ahead of their time, investors questioned the practical use, and it became a buzzword that serious people avoided. Of course, five years later that all started to change for AI. Well, here we are again. Apple Vision Pro proved to be niche. Meta’s audio glasses have garnered attention, but the impending AR glasses still seem lackluster, based on early reviews. While we have a very long final mile to go here, there is no doubt in my mind that the ultimate technology interface will be an inch away from our retinas and become a layer of insight, knowledge, and engagement on the physical world. Even in just the last few weeks, rumors around the potential of Google’s new glasses made in collaboration with Warby Parker are getting people excited…

How will we know AR has truly arrived and that the consumer product gold rush has begun? There is a common belief among consumer product leaders that, when it comes to a truly innovative consumer product, you can tell if someone is using it from across the room. In other words, truly breakthrough consumer products ignite breakthrough consumer behaviors. For example, when you see people standing and holding a phone at the edge of the street, you can guess they’re using Uber. People looking at their phone and smiling are probably taking a selfie on Snapchat, while people dancing in front of their phones are probably using TikTok.

I’ve often wondered whether you can reverse-engineer this system. Can new ideas for consumer behaviors breed new consumer products? During my morning run the other day around Central Park, I imagined a world where people said “hi” to strangers and were accustomed to being recognized and approached by people they didn’t know directly. I wondered, in a world where everyone wore AR glasses tightly coupled with a trusted social graph that illuminated an aura of information around anyone you were connected with, would it suddenly become interesting, if not polite, to say hello to friends-of-friends? If I’m excited to meet a fellow New Yorker on a random train in Tokyo, why wouldn’t I get a similar sensation from bumping into my friend’s old roommate whom I never met? Will new sensations of connection help restore the growing sense of loneliness in this world? I ultimately believe that the greatest consumer technologies take us back to the way things once were, an era we long for, but with more scale and efficiency. Truth is, we long to be known. I am interested in new technologies that will make the world feel a bit smaller for all of us.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few data points that caught my eye and have stayed on my mind as an investor, technologist, and product leader - as well as a preview of 10 predictions/outlooks for 2026 that are top of mind as a builder and investor. Subscriptions go toward organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) content creators being willing to trade control for speed – and artists definitively not, (2) a resurgence of hardware startups, (3) the perils of startups growing too fast and why some popular “unicorns” may prove to be rabbits, and (4) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • How will religion evolve in the era of AI? For the first time, we’ll get a response from something “all-knowing” about us. The transom with ASI (artificial super-intelligence) will work both ways. As AI works in ways that exceed our comprehension and becomes more mysterious, will its perspective carry weight? Will its advice become a self-fulfilling prophesy?

  • The final mile of leadership. Something happens in the final mile that truly tests leaders’ resolve and core values. When quick, key decisions can shift money around quickly, when the incentives of the leader and their team start to diverge, when the spotlight shifts, something changes. How a founder operates at such moments reveals so much. This is the mile of the journey by which leaders are truly measured.

  • Your time will become more valuable. I expect our culture to crave the exchange of undivided attention. Showing your hands during a zoom meeting, eye to eye contact at a table with your phone away, or voice memos and handwritten notes that demonstrate the time you gave to someone will have tremendous meaning. The sensation of giving someone your minutes will grow.

AI & Artistry

There are two types of people making stuff: content creators and artists.

Both are creative. But there is one distinct difference: Content creators are willing to trade control for speed. Artists are not.

• Content creators are willing to trade control for speed. Creators are trying to get a good-enough cut, a performant-enough ad, or an engaging-enough YouTube or TikTok post. Creators know that modern brands must think and act in real-time. They know they need to move fast and they understand that most timeline and algorithmically programmed content is ephemeral. Whether it is the pizza shop needing a promotional video or the influencer making an opening sequence for their gaming show on YouTube or creating social and meme-driven content for our own social feeds, we all play in this new creator economy. The advent of prompt-driven, generative AI tools lowers the floor so that anyone can be a creator of some kind. As a result, the market has exploded. Between 80 and 90 percent of the creative market is now focused on the creator: Canva, Adobe Express, MidJourney, ChatGPT, Nano Banana, Gemini, and every other mainstream generative creation tool caters to creators. But creators are driven by the desire to accomplish something quickly. The content is often a means, not an end.

• Artists are NOT willing to trade control for speed. Sure, artists do quick sketches and explorations, but these exercises are in service of exploring the full terrain of possibilities to find the absolute best direction. Artists have a distinct vision in their mind’s eye, and they will use whatever chisels, brushes, cameras, editing workflows, or other tools they need to achieve this vision with as little compromise as necessary. Artists require creative control and giving it up to some prompt-based generative tool is like the sculptor working without her hands. The artist’s creation, and achieving their exact vision, is the goal. Their tools must be in service of this goal. While a much smaller percentage of the creative market, artists make units of culture. Artists craft the meaning-infused stories that thrill and break us. While their work may not be fast or optimized for clicks, their stories teach us about ourselves and advance our understanding of the world.

• Content creators and artists have different needs. Most tech headlines about AI have been about the impact on content creators and the insatiable desire and demand for content of all kinds. The majority of venture capital has been invested in tools for the “creator economy” – making ads, making endless amounts of YouTube and TikTok content, making social posts, etc – given its vast, if not infinite size. But lost in the shuffle is what artists need. People conflate what enables creators with what might replace artists. People see the “AI slop” capabilities of new generative models and then declare the “end of fine art” and “RIP Hollywood,” as if the founding of McDonalds would kill our desire for great cuisine. It is increasingly clear that artistry is here to stay. We will all continue to crave great stories and breakthrough art despite the surge of AI slop, much like we crave fine cuisine despite the abundance of fast food. What do artists need? They need ways to explore more surface area of possibility more quickly so they can find the right idea to pursue. They need modern workflows across different tools that offer more precision than ever with fewer obstacles and dependencies. They need new technology to give them more creative control, not less. And artists must feel the appetite to take more creative risk because when they do, our lives become better. There is a new breed of workflows emerging for artists that leverage new technology without compromising control and I’m excited as ever for the implications.

While I’m excited about the creator economy (and suspect this will be a growing job sector, as opposed to many others), I firmly believe that we will all crave more artistry in the age of AI. As content creators fill our feeds and brands flood the zone — all to grab our attention as cheaply and efficiently as possible, consumer preferences will shift towards more crafted and deeply human stories. As creating content becomes easier, the bar for what makes an extraordinary and unforgettable story will rise.

The Resurgence of Hardware Moats

There’s a decent argument that, as we approach AGI over the next five-10 years, the foundational models will be capable and incentivized to compete with the capabilities of every piece of software that was built on top of them. Doing so opens up new markets to justify their capex spend and capture margin at the application level across every vertical that has been transformed by AI. In a world where any interface and application can simply be summoned and tailor-made, it is not hard to imagine mass disruption across software as we know it.

One implication will be the growing popularity of vertically integrated products that include hardware in the mix. The specialized hardware/software combo is an ultra-resilient moat, so long as you don’t run out of capital before your product is done! The companies that thrive will look like Whoop or Oura, where the AI is tightly coupled with a wearable; Board, where the deployment of digital games happens on their own collaborative, physical game board; or Meter, where a vertically integrated networking solution for the enterprise includes both the industrial-grade hardware as well as the AI-powered tools for networking engineers to deploy applications and manage an enterprise network.

The complexities of hardware startups have historically been a turn-off for investors, given the specialized talent, complicated supply chain, and enormous costs required to successfully develop, launch, and scale. But a new breed of “picks and shovels” companies has emerged that streamlines traditional obstacles like chip design and manufacturing, prototype development, and supply chain management. For example, Flux offers a full browser-based way to use AI to design chips that are ready for fabrication, whether its a connected device, robot, smart home device, drone, and the list goes on. Other companies like Lumafield are outfitting hardware engineering teams with capabilities that were once only available to the world’s largest manufacturing companies. As teams of all sizes are able to rapidly develop and iterate hardware, I suspect we will see more vertically integrated hardware/software startups emerge that cannot so easily be replaced by an ever-powerful frontier AI model.

Are most unicorns actually rabbits?

Regular readers of Implications know about my obsession with running, but I have yet to run a marathon. If I ever do, I would pace myself to reach the finish line, as opposed to being one of the many “rabbits.” In marathon terminology, a rabbit is someone who starts a marathon running super-fast and, for a short while, is “winning” before quickly falling behind — unable to sustain the breakneck pace.

As the AI platform shift hits, we’re seeing a whole series of companies run fast out of the gate, in the form of record-setting growth rates and tens, if not hundreds, of millions in revenue within a couple years of their launch. But when you dig in, you learn that they’re not yet economically viable. And, in many cases, they are surprisingly interchangeable with competitors when it comes to their core utility. Rabbits are winning, but unsustainably. In some cases, they are selling their product for much less than the cost of the underlying API calls with LLMs. Anyone who sells a dollar for ten cents is going to have record-setting growth and a ton of revenue, but they won’t last.

Even before the AI era, as we look back at many of the so-called “unicorn” companies (those that raised money at a $1B+ valuation) over the last decade, we’re realizing that many of them were rabbits. Many of these companies leveraged venture capital to acquire customers without understanding (or willfully ignoring) lifetime value or unit economics. Some of these companies told a story in the press that misrepresented the true nature of their traction. Perhaps we should have known better when hundreds of unicorns were minted. If unicorns are everywhere you look, are they still unicorns?

How does a high-growth, ambitious company avoid becoming a rabbit? Seek a valuation that makes your newest employees feel rewarded over the next two years (as opposed to seeing their equity value stay stagnant or lose value in a down round). Challenge yourself to improve customer acquisition costs in tandem with growth rates. And focus on customer impact, especially at this moment in the platform shift in which products hit huge growth numbers (everyone wants to try everything!) but struggle with sustained use and retention. Set goals and celebrate milestones that are more related to customer impact than wild growth and top-line revenue without any path to profit.

What makes a true, rare, special, and one-of-a-kind unicorn? In my book, if a company achieves $50M+ in revenue with a sustained, if not growing, 40+ percent growth rate and line of sight to achieving profitability, if not already cashflow-positive, it is remarkable. Otherwise, if a company is making tons of revenue and raises at a $1B+ valuation without such characteristics, it is liable to be a rabbit.

Portfolio highlights & opportunities

A few updates on companies i’ve supported as an angel:

• Mem0 launched their universal, self‑improving memory layer for LLM applications, powering personalized AI experiences that cut costs and enhance user delight. The Mem0 team has been hustling for quite some time to help companies remember their customers in the age of AI — rather than have this knowledge live with their foundational model providers. The team is hiring engineers and dev relations positions.

• Lightfield had their public launch of “an AI-native CRM that remembers everything, then acts on it. Lightfield automatically captures and organizes every customer conversation across email, meetings, and notes, creating your team’s collective customer memory. Lightfield helps founders learn from every interaction, which has a ton of unexpected implications when you start to search or pose queries.

• A product called “Shelf” from Koodos has been growing on all cylinders as a way to automatically show the music, shows, books, and other things you’ve enjoyed this week to friends. I highly recommend playing with Shelf. Under the hood, Koodos is building a graph of user preferences that can help us discover things we want to do, people we want to meet, and more.

• Knock, a company that makes infrastructure for sending product and customer messages, released Guides, which has had overwhelmingly positive customer response. Guides lets you use your own components (rather than slow external JavaScript libraries) to power in-app messages that drive activation and upsells. If you’re building an app and seek world-class notifications and messaging for your customers, I highly suggest checking out Knock.

• Scribe recently announced their Series C round of financing at a $1.3B valuation, which is notable given how under-the-radar this team has been since their initial launch. They started by building a standard for documenting how work gets done: self-writing documentation that captures invisible know-how and scales it to colleagues, customers, and clients. They have since launched AI capabilities that are getting customers excited. The team is hiring for a Director of Talent, production management, and more.

Ideas, missives & mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including a perspective on heritage as a moat for brands, the three layers of a modern story, learnings from Coca-Cola on the perils of using AI without creative control, and a few data provocations). Subscriptions go toward organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) the growing ubiquity of data and what new forms of data may prove valuable, (2) the collapse of social, sales, and support functions into one — and the perils of seat-based business models, (3) why generative-only social apps will struggle with retention, and (4) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • While people with less tech experience are especially vulnerable to the daily occurrence of scams, all of us are facing these tricks and puzzles these days and the precision and velocity of these crimes will only grow. We need a new “safety layer” enabled by AI.

  • I’ve long had concerns about the impact of attention-optimized algorithms on social media. Once you understand how these algorithms are designed, it is easier to understand why they reinforce our strongest views and actually hide any nuanced and more reasonable “both sides of the argument” content (agreeable and reasonable content doesn’t perform, and thus never gets displayed).

  • More recently, I’ve been thinking about the loss of nuance inspired in part by a Palestinian friend and fellow builder. In a recent conversation, it was especially clear how much nuance is lost in the public narrative, thanks to the algorithms. The business models of journalism simply do not support nuanced reporting. The primitive part of our brain fears it. And the algorithms altogether hide it.

  • Consumer-facing technologies like ChatGPT and Claude are extending their memory of our conversations and it is only a matter of time before we’re given the option to share the AI’s memory of us with our colleagues, family and loved ones.

Will everyone’s data be rare or everywhere? What’s the next frontier of valuable data? Graphs? Robots’ memories?

We’ve discussed data in previous editions of IMPLICATIONS, and went deep on the data wars and how to navigate building products amidst a platform shift. But is the underlying assumption – that data is the precious moat for the era of AI – even true anymore? If you’re a developer, or paying attention to any of the new AI startup onboarding experiences, you’ll notice that every company is trying to sync everyone’s data. New companies in the enterprise search and agent space often launch with dozens if not hundreds of “connectors” that enable you to instantly query and index real-time data from other apps. And while we chronicled the crackdown on this by some major players, there’s a decent argument that personal and corporate data will become the most transferable thing in an era where any self-respecting company will have a copy of all their data under their own control via their own Snowflake or Databricks instance. So, when historical data is no longer a moat, what will be?

• Graphs will become more valuable and proprietary. Whether it is a “permissioning graph” in a company (who is allowed to access what), a “teamwork graph” (the proprietary understanding of who works with whom, project by project, much like the graph Atlassian has constructed across all of its “system of work” products), or a “social graph” (the map of who your friends are, and the density of your relationships), these graphs take many years to build and cannot be easily copied and ported. As a result, these graphs will become more valuable and serve as moats in some industries.

• Portable personalization profiles will become the preferred login option. As we’ve discussed in previous editions like “The Personalization Wave,” the future of search, commerce, and decision-making will be hyper-personalized through interfaces that cater to your style and preferences. In short, personalization effects are the new network effects. As we interact and do more of our everyday thinking alongside LLMs, they will get to know us better than we know ourselves. This “memory” of us becomes an instant and portable force of personalization. If you played with the new Sora social video app, you’ll notice the default “Sign in with ChatGPT” option, and you can quickly imagine the implications of this becoming an option on any website or e-commerce brand you visit. The next generation of the web and your everyday experiences with software will revolve around YOU, powered by portable personalization profiles.

• Raw and real-time environmental data sources will become more valuable, and increasingly sourced from robots. What happens when AI has leveraged the internet’s data and essentially runs out of data to train on? What happens when our cars, devices (glasses?), and the next era of autonomous machines require an entirely different form of real-time environmental data? The greatest new real-time data source that you won’t find on the internet is coming from robots. Whether it is a vacuum cleaner that intricately maps your home (like Matic, which seems to be all the rage in Silicon Valley these days), delivery bots and drones, or weather detection buoys and robotic traffic cops, these roaming machines that permeate our lives will offer a bounty of data that is unlike any other. My bet is that data captured by robots will greatly outweigh and outperform data captured from the internet for training the next generation of industrial grade models. One idea for readers: Another fun dataset to be developed and used for AI training is games between humans and robots, where the nature of human thinking can be better understood. Anyone doing this?

• Long-tail specialized and unspoiled data sets will unlock precision automation. Another type of data you won’t find on the internet comes from very purposeful sources with meticulous collections, like atmospheric data from companies like Atmo.ai. Original nuanced datasets and models from companies like Roboflow will also become more valuable. A recent catchup with Joseph Nelson, the founder/ceo of Roboflow, further illuminated the utility of very narrow data sets in the age of AI. As he explained to me, whether it is identifying a particular nuance of an assembly line with a vision model or leveraging a model trained on the workflows of a warehouse to improve worker safety, the opportunity to capture very specific data in the wild to build narrow AI models is enormous.

The AI era will mash enterprise verticals together – and the products will follow suit.

In edition #26 of IMPLICATIONS, we discussed how different stacks of products and people in companies of all sizes are increasingly becoming one. In particular, we’re seeing increasing overlap between tools devoted to functions like support (like Intercom, Zendesk, etc), tools for social media management (like SproutSocial and Hubspot), and new “sales agent” tools that help convert customer inquiries into sales. If you squint, what’s the difference between a customer seeking help over social vs. over email? What’s the difference between a customer asking a question about a product vs. considering purchasing a product? What are the implications?

• Customer paths will be disrupted, causing the collapse of company functions into one another. In most big companies, social accounts are still managed within the marketing organization, the support function is often outsourced and managed with a completely different technology and set of goals, and sales stands alone as its own team with a distinct tech stack. And yet, the customer has stopped discerning between these paths. Even within an enterprise market, customers do their product research on LinkedIn and seek support via ChatGPT. Organizational design and technology solutions are now living in a different era than the customer. As a response, new companies will increasingly centralize social, sales, and support as modern “customer success” organizations leveraging a new breed of tools and customer-facing agents that collapse these channels and goals into one another. This also means that “support” may shift from being a malnourished cost center of most companies to a well-resourced and consistently optimized profit center that grows the business.

  

• Sales organizations will split as outcome-based products challenge seat-based pricing. In edition #11 of IMPLICATIONS (back in 2023!) we discussed the strange ways AI could disrupt business models, but one of the most threatened business models of the pre-AI, human knowledge worker era is seat-based pricing. Seat-based pricing is entirely predicated on the fact that a single human can only accomplish so much, causing the need to hire more humans and pay for more seats as a business scales. In such an era, sales incentives were aligned with the industry – companies added more employees and salespeople sold incremental seats. Now, as individual employees can spin up countless agents to scale, salespeople pitching more seats is sounding increasingly tone-deaf. Modern companies will increasingly aspire for talent density in the form of a smaller group of more productive and more capable people, as opposed to “growing seats” as they scale. AI-native startups are taking advantage of this new world with outcome-based pricing (e.g. pay per new lead, per new project, per item sold, etc) with a simple “let us reduce your seats of X” sales pitch. I anticipate smart incumbents will eventually cannibalize themselves by splitting their sales teams, having one team sell seats and the other team sell against seats to those customers who are ready. Won’t be fun, but will be necessary.

Generative-only social apps will fail to sustain engagement without the human fingerprint.

There’s a lot of excitement about new consumer apps in the age of AI, and we’re seeing some smart contenders with Sora’s video app among others. I am excited about these explorations, but my somewhat contrarian take is that generative-only social apps will fail to sustain engagement unless they incorporate a truly novel human element. Why are they likely to fail and what tactics might work?

• There are three reasons new social or media apps with entirely generated content (content made solely from prompts) are likely to fail: (1) the “ego analytics” that cause people to come back and see who engaged with their content are missing when AI made the content. Without a mechanic to make creators feel great about their content, there will be no obsession and therefore no retention, (2) the lack of human craft and taste required to make the content will translate into disinterest among the consumers of the content. Without a human story behind the story, will humans care as much about the content they are watching? (3) the ease of creation will accentuate the shallowness of the subject matter. As creators, we have a higher bar for what we make when it requires friction to make something. But if it’s super-easy to make something, there is no bar.

• New creator technology is only enticing if it unlocks creative risk-taking and human storytelling. Don’t get me wrong, I am so bullish about the evolution of creative technology and firmly believe the future of creative control and highly crafted storytelling will include all sorts of AI-powered tools, alongside 3D gaming engines, world models, as well as traditional CGI and VFX technologies (all of which were also originally shunned before they were celebrated with technical Oscars!). But the evolution of technology doesn’t always change what humans desire. We still want unforgettable stories and crave the story behind the story as a way to relate. Whether subconsciously or not, we value art by understanding who made it, the history of the artist or actor or filmmaker, and what went into it. We must feel it is as real as possible to relate. Why do people crave human fingerprints in the art and stories they consume (regardless of the tools used to create it)? Because we are human, after all. And art has always been a path to understanding ourselves.

Portfolio Highlights & Opportunities

• Tolan: If you’re excited about the world of digital companions that remember you and are getting real traction in the market, you should definitely check out Tolan in the app store. I’ve had a ton of fun playing with it. The team is actively hiring a Designer and an iOS Product Engineer.

• Macroscope: Launched by longtime friends Kayvon and Joe (with whom I worked with closely during their last company, Periscope, before their Twitter acquisition) have built a remarkable product that helps teams understand their codebase and answer the elusive question “what did we actually do today?” As a leader of an engineering team, I have found the product immensely helpful - highly recommended.

• Symbiotic Labs: One of the demos in the “next generation advertising” space that blew me away was from Symbiotic Labs, founded by a group of deep thinkers and researchers from technology and psychology. They’re searching for their founding product design leader, and I know many investors are excited about what they’re building.

• Bubble: One of the original app development platforms that helps entrepreneurs and teams build their own software has recently launched a ton of AI capabilities that bring DIY software to another level. The team is looking for a few key engineering leaders, namely VP, Engineering , Engineering Manager, Platform , and a Senior Software Engineer, Mobile .

• Doctronic: A relatively new investment that is providing an AI-assisted but doctor-overseen first mile experience for healthcare needs. The team is growing on all cylinders and needs more senior engineers and customer success specialists, especially those based in NYC.

• Braintrust Data: Another company growing on all cylinders, Braintrust is the evals and observability platform for building reliable AI agents. As agents enter the workflows of our daily habits and systems of work, Braintrust enables teams to observe and improve. They are hiring across software engineers, GTM, and legal.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including the ongoing debate about a dollar debasement, unemployment trends, always being willing to bet your company, and some important data provocations). Subscriptions go toward organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) new technology for creativity without compromising control, (2) the tensions between nuance and tribalism, (3) the final mile measure of a leader, and (4) some surprises at the end, as always.

• If you’re new to IMPLICATIONS, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and subscribers. I aim for quality, density, and provocation rather than frequency and trendiness. We don’t cover news; we explore the implications of what’s happening (and some essential yet uncommon insights for makers). My goal is to ignite discussion and socialize edges that may someday become the center.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • As we reclaim more time, we’ll value it even more. When minutes become more valuable, what cultural practices and norms might change?

  • In an increasingly automated world managed by technology that optimizes for efficiency, the unsaid reasons we do what we do will become critical insights for product leaders.

  • We’re just years away from being able to share our memory (in the form of all of our interactions with AI) with colleagues, friends, and family. What is it like to access parts of the accumulated knowledge and experiences of others just as you recall your own lessons learned? What are the implications?

Alright, let’s dive into Edition #35 of Implications…

Empowering Creators vs. Refactoring a Process

New technology for creatives is notoriously misunderstood, if not feared, at first. From products like Photoshop to digital cameras and green screens, new methods are typically shunned before they are celebrated. Why? Because the process behind art is part of the art. The process influences how art makes us feel, and the best process innovations unlock new forms of art. For instance, the emergence of indie film was largely made economically feasible by digital cameras. The goal of new technology shouldn’t be to improve a process but rather to empower creators to do new things.

I’ve spent a lot of my career building platforms and tools to help creators make ideas happen. While people get excited about new technology, a common mistake is focusing on what the tech can make, rather than how it gives creators more control over what they make. The breakthrough happens when new tech empowers rather than replaces. The technology that actually extends creative possibilities (rather than simply refactoring work) is technology that helps creators take more creative risk to achieve what’s in their imaginations.

As we enter the AI era, I am seeing a lot of tools attempt to take away creative control rather than enable it. Many of these tools, albeit still in their infancy, are only able to satisfy those who don’t know exactly what they want, as opposed to the creators who do. The latter are always let down by inconsistencies, a lack of control over detail, and the frustrating game of adjusting one thing that inadvertently changes another, like a game of whack-a-mole. These new tools are attempting to replace process as opposed to enabling new (and more) possibilities. Asking a creator to change their workflow for the purpose of accommodating a new technology (to save costs, to be an early adopter, or otherwise) is like asking someone to do their best work with one hand behind their back. Tech for tech's sake is a fool's errand.

The implications? In the age of AI, I am increasingly convinced that human creativity is the craft of making something specific — something you see in your mind’s eye — and creating with more intention than ever before. Prompts will not suffice. For makers of creative tools and stewards of creative process, the real opportunity is to make tools that are ultimately in service of the story - of the creator’s vision - as opposed to simply refactoring or changing a process because you can. Your job, as an outfitter of creators, is to empower people to make new ideas happen. You’re enabling creators, not changing them. New technology should be used to solve creative problems in better ways, explore more surface area of possibility to find better solutions, and achieve more specificity. Too many new tools ask artists to compromise the details of their vision in return for having the technology do part of their job. True artists don’t want that deal. Creative risk is the source of unforgettable stories and breakthrough projects that move the world forward, and technology should help this happen. Every tool must be used in pursuit of whatever’s in the mind’s eye. Otherwise, it is constraining the story.

How Nuance is Lost in Tribalism

I was recently sitting down with a Palestinian friend and early-stage founder I deeply respect. His family were refugees and, through a twist of fate many years ago, he was the recipient of a fully subsidized scholarship from a European organization that brought together a cohort of Israeli and Palestinian children to study together at boarding school. While his extended family initially objected to him “being indoctrinated by the west,” his parents ultimately supported him accepting the scholarship, if only out of concern for his safety at home. As a cohort traversing their teenage years together, they all became friends, dated each other, attended each other’s life events, and emerged as young professionals with an enlightened perspective on their regional conflict and cultural differences.

I was interested in what his cohort of friends make of the current middle east chaos and information wars at work. It seems the consensus is a desperate hope for peace and a deep disdain for governmental leaders on both sides. But most of all, his cohort of friends have a deep appreciation for something that everyone else seems to miss these days: nuance. Nuance is the collection of caveats and footnotes beyond the headlines that one can only fully understand by being on the ground and adjacent to those suffering loss and hopelessness. Why have we lost nuance? Because, in times of danger and strife, humanity defaults to tribalism.

In the early days of humanity, tribes were our only way to be safe. What was once a literal group of people dedicated to each other’s survival — keeping people fed and allowing some to rest while others patrolled — is now a metaphorical group of people that are committed to reinforcing the world view of each other. Effective tribalism requires loyalty to the tribe and practicing and reinforcing whatever the tribe believes, often dictated by its leaders. Tribes must keep things simple to keep people together: you’re either with us or against us. And there is no room for nuance in tribalism. The lowest common denominator of belief systems is a clear understanding of what’s right and wrong. Nuance is a lethal crack to tribalism, because it allows doubt to creep in.

When facing the danger of a lion lurking, there is no safer place to stand than the middle of a tribe, surrounded by those committed to protecting you. The author Sebastian Junger wrote a great book making the case that, in the modern day, humanity still defaults to seeking safety in tribes. We are all biologically programmed to seek being part of a group and to feel safer in the middle of that group. Standing in isolation was dangerous in the early days of humanity – and remains dangerous today. When you’re not a part of the right or the left, who are you? What do you even stand for? Attempting to understand nuance and rejecting a tribal “right or wrong” stance requires bravery these days.

We’re getting more tribal these days and, as a consequence, losing our search and appreciation for nuance. To make matters worse, technology is making it harder to access nuance. As we increasingly rely on social media for our news, powerful algorithms tune out nuance because it doesn’t engage us. We favor sensational and declarative headlines rather than fair, nuanced takes on a situation, and the algorithms optimize accordingly. The business models of journalism simply do not support nuanced reporting. The primitive part of our brains fears it. And the algorithms altogether hide it. As a result, tribalism is further fortified on all sides of an issue.

The True Measure of a Leader is Revealed in the Final Mile

Lately, as talent wars in the tech world lead to acquisitions that disproportionately favor the founders, there has been much discussion about a founder’s responsibility to their team in an acquisition. In my experience, founders know, in their gut, what the right thing to do is. Ultimately you should do what you want to remember doing 10 years later when you look every team member in the eye (or try to hire them again!). The dirty little secret about leadership is that a large amount of everyday leadership is performative. When things are going as planned, many leaders are capable of all the best practices for team building, strengthening a culture, and driving performance.

But I’ve seen enough leaders behind closed doors during perilous or highly consequential moments to realize how much of a leader’s impact comes down to decisions made and actions taken in “the final mile.” Whether it is the final period before a product’s launch, the lead up to an IPO or acquisition, or the final period before a company is shut down…nothing reveals a leader’s true character and capabilities more than this final mile. Most leaders try to do the right thing on a daily basis to ultimately serve customers and their teams. But something happens in the final mile that truly tests leaders’ resolve and core values. When quick, key decisions can shift money around quickly, when the incentives of the leader and their team start to diverge, when the spotlight shifts, something changes. For instance, many leaders overestimate their own importance and undervalue their team in the final mile when they negotiate deal structure and determine salaries and titles. No matter what you hear, the truth is that vesting can be accelerated, cap charts can be tweaked, and founders are often empowered to do the right thing for their teams (and ultimately for their own pride and reputation in the long term)…if they make it a priority. How a founder operates at such moments reveals so much. This is the mile of the journey by which leaders are truly measured.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. **We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including how stories are the building blocks of culture, the bundling cycle of media, and data provocations around the economy, how we use LLMs, and and a few other curiosities). Subscriptions go toward organizations I support including COOP Careers and the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) the unlock of your precious time and attention - and what may transpire as a result, (2) the sway of brand in the AI era - an argument for and against the importance of brand, and (3) some surprises at the end, as always.

• If you’re new to IMPLICATIONS, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a group of now ~40k subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help you connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • How might AI unleash an entirely different next generation of social apps, media and entertainment platforms, and other consumer-oriented experiences? Our entire future cannot be limited to an endless series of chats!

  • Modern tools that unlock 100x more cycles of ideation and concepting are making a profound impact on the creators who use them. Here are some thoughts on why the future of creativity is bright, albeit different.

  • There is much focus on how bad actors may use AI to trick us (fake media, cloned voices, extended context windows to socialize us for a long-game scam, etc), but how might AI ultimately save us?

Your most precious resource (and the ultimate gift) in the era of AI is your human time and attention.

“AI will free up so much time” is a common claim these days. But if you really think about the implications of being infused with superpowers to achieve more per minute, it gets a bit more complicated. Hear me out…

• As we reclaim more time, we’ll value it even more. This is somewhat paradoxical, but it’s important. In a world in which our energy and potential is compounded by the technologies we use, we will all be able to achieve more per minute. You’ll offload all sorts of tasks and explorations to agents, run countless automated tasks at once, and have logic-driven technology make decisions on your behalf while you sleep. As it becomes possible to do more every minute, a minute becomes more powerful. You might think this means that you’ll be more free with your newfound “spare time” minutes you reclaim. However, “Jevons Paradox” suggests the opposite. Jevons Paradox applies to circumstances where increased efficiency of a precious resource leads to a rise in the total consumption of that resource. In many situations, including the use and availability of semiconductors and chips for computation, and our own most personal computers (our brains), gains in efficiency do NOT lead to conservation. Instead, as these resources become more efficient, they become used even more than ever before. Perhaps, as our minutes become exponentially more powerful, we will want to do even more with them? My bet is that we end up wanting more minutes than ever before as the ROI of a minute goes up. In short, your minutes will become far more valuable to you (and to your employer, now that you are so much more capable).

• When minutes become more valuable, what cultural practices and norms might change? Given time is truly the most precious commodity (and will be reclaimed as well as grow in value), my bet is that we will spend our time with more intention on scarce things that AI and tech cannot do for us. For example, we will crave human-crafted experiences, like non-scalable chef creations, artistic journeys, and travel off the beaten path. Things that require time will become more valuable. I hope this trend ignites a new wave of small businesses that create such experiences as the labor markets shift to accommodate AI and the drastic cut in knowledge worker jobs. I also wonder how, in our personal lives, we will start spending our time differently as it becomes more valuable? Will we make more things by hand — or by mind — that could otherwise be automated, if only to infuse time (value) into these things?

• Undivided focus and your precious human time are among the most scarce and meaningful gifts you can give others in the AI era. Especially as algorithms becoming more sophisticated in competing for our attention, giving your time to a human will be a gift. I expect our culture to crave the exchange of undivided attention. Showing your hands during a zoom meeting, eye to eye contact at a table with your phone away, or voice memos and handwritten notes that demonstrate the time you gave to someone will have tremendous meaning. The sensation of giving someone your minutes will grow. Finally, I think the bar for entertainment goes up dramatically as we become more protective of our attention and look for brands to signal the types of deeply meaningful stories that deserve are ever precious time. This trend was certainly a driver for me to join A24, and I think it will play out as top-tier stories become further differentiated from the masses of content.

The sway of brand in the age of AI: What happens to “brand” when we outsource our purchase decisions?

A common debate I have with leaders in tech and consumer products is whether “brand” will be more or less important in the AI era. One side argues that popular opinions about brand — and the general sway of a brand — will become less important as so many of our purchase decisions are made by a hyper-personalized AI that knows our preferences and past purchases so deeply that we just trust it over the influence of any brand. The other side of the argument is that we will rely even further on brand because of the rapid commoditization of everything, and the fact that LLMs will present options in ways that make options appear so equivalent that “brand” may ultimately drive our choices even more. There is also a great case to be made that, as humans, our natural tendency when something becomes abundant is to seek scarcity and meaning-infused versions of everything, signified by brand. Let’s explore some of the edges and key technologies helping determine the future influence of brand and our buying behaviors…

• Buying Agents for consumers: What happens when we outsource our purchase decisions to AI? Transactional behavior will change dramatically in the age of AI. The first phase of this era is simply relying on ChatGPT answers, but I anticipate a world where agents that know us extremely well (including our dietary and fashion preferences, sizes, current closet and pantry inventory, etc) will become a very trusted if not default decision maker for our purchases. If you really think about it, why would you ever trust a company’s heavily biased product descriptions and determinations on what to merchandise to you when you can summon the knowledge of humanity coupled with a highly trained agent that is deeply attuned to your likes and interests to suggest the best options? Also, will you also depend on your agent to advise you not to buy? Perhaps budget-driven advise like, “do you really need another guitar when it will delay buying a house for another month?” Or data-driven guidance like, “this is a really bad time to buy a flight to Italy — better to wait until fall when it’s cheaper and less crowded.” Of course, these are just a few more of the many wild changes to be expected for the future of commerce.

• Consumer brand agents and support agents begin to blend. I have seen a growing number of startups and established marketing technology companies begin development of a new variety of agents that work on behalf of a brand to engage consumers and personalize shopping experiences. “Hi Scott, welcome to On Running, how can I help you? Are you looking for running shoes today or something else?” It is clear that we will soon be greeted by digital agents on every website we visit. Some will be audio-first experiences, some will remember our past history with the brand, while others will request access to our broader “memory” stored with an LLM like ChatGPT or perhaps a platform-agnostic player like Mem0, a company from my seed portfolio getting a lot of traction these days. But big questions loom: Will these consumer brand agents be connected to a company’s marketing stack? Will these agents also be designed to handle customer support inquiries? Will they only operate within a company’s website, or will they integrate into our general AI tools or perhaps even other brand’s websites via some form of partnership? Are they considered marketing or support tools? Does it even matter?

• Enterprise brand agents and sales agents begin to blend. I suspect we are VERY close to being greeted by an enterprise sales agent on every business-to-business services website we encounter. But one of the most exciting disruptions here could live within the procurement process itself, as a new breed of procurement agents (perhaps powered by market disrupters like Globality, a company I know well) take your order or request via a conversational form of an RFP, like “evaluate CRM options for teams of twenty, and propose the best value and solution for my company,” and then do all the research, run price and feature comparisons, perform reference checking, and perhaps even the negotiate on your behalf. Rather than search Google, ask peers in the industry, and consult industry analyst reports, you’ll start tasking agents to do all the research and make an objective recommendation. What are the implications? First of all, the industry of influencing LLMs for such activities is going to be massive (see next bullet point). This world will also give rise to specialty buying agents that are trusted and trained on a very structured and controlled dataset.

• Brands must be optimized for an era in which AI recommendations will outperform human discovery. The emerging field of “AI visibility” and tailor-made tools to either track your AI visibility (where and how AI mentions your brand) as well as optimizing your brand and content for AI are becoming an absolutely critical part of every marketing stack. I’ve gotten to know the leader of this market, a company called Profound, since their seed stage and their co-founder James and team have helped me understand some of the implications for brands in the age of AI and the criticality of boosting your product visibility within shopping-oriented queries in ChatGPT, among other mainstream LLMS.

• Ultimately, “meaning” and human-crafted story will be the ultimate differentiator. For the last few years, we’ve explored in IMPLICATIONS the concept of a “meaning economy” in a world of content abundance and zero-cost content creation. What is meaning? It is the story, the purpose, the myths behind the creators, brands, and objects themselves, and the differentiating brand value that results. When anything becomes commoditized or ubiquitous — whether it is shoes, a popular restaurant that becomes a chain or content that gets reproduced endlessly by some soulless AI app — consumers tend to crave a more scarce, special, and differentiated alternative. What makes something scarce and differentiated? Meaning. As I often remind teams getting excited about the use of new AI tools for creativity, the science of business is scaling, but the art of business is the things that don’t scale. In the meaning economy, the art part becomes increasingly critical. As artists spend less time doing mundane repetitive stuff (thanks to AI), they’ll have more time to explore more ideas, take more creative risk, and ultimately create better stories.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader including surprising trends in the gap between popular and highly rated content, a contrarian view on clunky incumbent enterprise software, thoughts on flexibility, and a few data provocations and graphs that remained on my mind. Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) the rise of a new and wildly imaginative crop of consumer apps in the age of AI, (2) the unsaid reasons we use products, (3) uncommon yet critical practices for leading within big companies, and (4) some surprises at the end, as always.

• If you’re new to IMPLICATIONS, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and subscribers. I aim for quality, density, and provocation rather than frequency and trendiness. We don’t cover news; we explore the implications of what’s happening (and some essential yet uncommon insights for makers). My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • There is much focus on how AI will make things worse (fake media, cloned voices, extended context windows to socialize us for a long-game scam, etc), but how might AI “safety layers” save us?

  • The future of commerce is going to look wildly different, with micro-payments on our behalf, new “on the fly UIs (user interfaces)” that simplify choices, and hyper-personalized pricing.

  • At work, colleagues will be able to mine each other’s interactions and realizations, and network in unimaginable ways to advance a business. This phenomenon is called “collective memory.”

  • The ongoing battle to be the top interface layer continues. I call these “the data wars,” and they seem to be playing out since we first discussed it in April. For consumers, the battle will be fought primarily across operating systems as well as emerging AI-powered browsers and immersive experiences. At work, the battle will be fought across every function.

• Alright, let’s dive into Edition #33 of Implications…

The missing topic: Consumer AI

I have been to two major private tech conferences over the last four weeks, featuring some of the greatest investors in tech discussing trends and their prognostications about the future. And the question I’ve been asking myself after each conversation and presentation: What is nobody talking about yet, but will be?

I have long believed that the greatest investors have a deep understanding of the present, but often struggle with the future – to fully appreciate the edges that may become the center. Whenever I go to these events, I challenge myself to come up with one theme of interest that nobody seems to be talking about. At this moment, the missing topic is “consumer AI” beyond the skeuomorphic natural language “chat with an LLM” products that we have today. How might AI unleash an entirely different next generation of social apps, media and entertainment platforms, and other consumer-oriented experiences? Our entire future cannot be limited to an endless series of chats! The consumer tech world needs to get more interesting! Let’s explore some of these potential edges of consumer AI that might seem crazy at first, but could change our lives over time.

• Peanut Gallery: I shared previously a consumer app idea I was brewing (and if you want to build it, contact me!) called “Peanut Gallery.” Imagine a simple social app where you post a piece of media, and then you get dozens of witty, funny, informative, and supportive responses from a collection of 100 AI-powered characters that each have a very specific backstory and set of characteristics. Some of these characters can be historical figures like Ayn Rand (characters that are now in public domain). Some of these characters can be completely made-up personalities. Perhaps some of them are sanctioned digital twins operating on behalf of your friends, with some degree of plausible deniability? These AI “people” comment on your posts, and each other’s comments. Sometimes they fight with each other. ONLY AI-powered characters are allowed to comment. The core novelty of the product is seeing all of the wild and hilarious AI-powered responses to your content and your friends’ content. The responses roll in over 72 hours after every post (so users are constantly tempted to login and see the latest). As the poster, you can also respond to the responses and start hilarious conversations w/ these AI personas as a public spectacle for your friends and others to watch and see. Philosophically, this is a way for humans to learn about AI, its limitations, and play with the idea of humans starting the conversation and seeing where the machines take it. Perhaps we use such environments to trial new posts and ideas before sharing with humans? Bigger picture, this idea introduces mechanics where we (humans) are the spectators, and typical social boundaries can be challenged when our autonomous digital representatives are able to act on their own accord.

• Social Simulations: Imagine deploying your digital representatives for all sorts of social explorations. For example, not sure how to play out a situation with friends in your high school? Debating how to approach your girlfriend or spouse regarding a particular issue in your relationship? We’re not too far from being able to deploy a digital twin of yourself (based on your own AI’s supreme and comprehensive understanding of you and your tendencies) alongside the digital twins of these other people — with their distinct personalities and full history of online behavior and other characteristics (and hopefully with their consent or participation!?) — into a simulation alongside your own to have a “sit back and watch” experience of the various scenarios that may possibly ensue. Will it someday be irresponsible to make a personal life decision without running a simulation first? After all, we try to do this mentally today with limited success based on our inexperience and limited skills in doing so. The same applies in our professional lives — can you imagine having a conversation with a digital twin of a customer or a boss, to rehearse and predict possible outcomes? I can see both novelty and utility in these possibilities.

• Alien Friends With Deep Memory: I have been having a lot of fun playing with my Tolan, a remarkable and fast-growing social consumer product designed to help users feel grounded, inspired, and connected (and recently part of the capitalB portfolio). The product starts as a cute alien companion to riff off your thoughts and help you make sense of anything you share with it. Early users rave about the ideas and feedback their Tolan provides them, and I find the engagement patterns and unexpected use-cases fascinating.

• AI Dating & Wingman/person as a service. Imagine the perfect sidekick that chimes in to tell new friends or potential first dates things you would never say about yourself - how amazing you are, some facts about your empathy and passions, etc. To be credible, your sidekick would also share your vulnerabilities (but in a way that solicits empathy!). While I’m not single, my younger, single friends (who are actually looking for relationships) say that they want dating tech to change from being a higher frequency less-vetted top of funnel to a lower frequency but more highly informed group of vetted matches. I wonder how AI can help make matches with a higher dose of transparency (explainability to open minds on topics like WHY certain opposites attract, etc)? Also, can AI somehow facilitate interactions between people? Can a prospective date’s AI agent meet yours before the introduction is even made? Perhaps the interaction between two digital twins can become fun conversation fodder for your first date? You’ll come in knowing far more, and be able to determine whether there’s a match far more quickly as a result. I suspect similar apps may emerge for general relationship building. As an introvert, I have always found the first mile of relationship building to be particularly straining and I would welcome any way to offload small talk!

Respecting the unsaid reasons for using products.

As the world becomes increasingly automated for the sake of improving efficiency and removing the many frictions of daily work and life, it is tempting to dismiss the future need for many roles, professions, and products. But one humbling realization I have had across all the products I have worked with is that “people aren’t rational.” In other words, we do things in our lives — and use certain products and services — for unsaid reasons that are extremely important to us personally, even if they cannot be described to others.

I recall a product in the enterprise accounting space that got far more utilization once its end users were able to export their reports as editable files (so they could put on their name and logo before sharing it more broadly). In other words, the unsaid reason they used the product was to get credit for their work.

With dozens of physical fitness apps, virtual AI trainers, and YouTube videos at our disposal, there is less rational reason for paying a personal trainer to run you through a regular training session. Nevertheless, I use a personal trainer (whom I love, hi Jaynee!) because she holds me accountable for showing up – my relationship with a professional human who has known me for years drives my commitment and performance. In this instance, the unsaid reason is my own accountability and commitment, not the workout instructions on their own.

When it comes to consumer products like social media, the dirty little secret is that we login more frequently AFTER we post content, so we can see who liked our content. In other words, we think we use Instagram to see what other people post, but the unsaid reason we engage is our own ego.

To succeed, technology that seeks to remove the friction of our lives (or accomplish tasks with supreme efficiency) must also recognize the unsaid, sometimes irrational reasons we actually use an app or service. These are the unsaid reasons we do what we do, and in an increasingly automated world managed by technology that optimizes for efficiency, the unsaid reasons we do what we do will become critical insights for product leaders.

Uncommon practices for driving innovation within a big company. (Part 1)

Having spent time as a founder (Behance over seven years, and now again at A24 Labs), an investor in many startups, and a chief product officer at a large public company, I’ve become rather skeptical of many conventional leadership best practices. Over the years, there are several somewhat contrarian practices for driving innovation that I have observed among others I admire, that I’ve attempted (sometimes but not always successfully) and that I firmly believe in.

• Port people and products to benefit the company and block fiefdom building. It is rare for leaders to proactively transfer great people doing great work somewhere else in the company. After all, “don’t fix it if it ain’t broken” is the default in a busy business, and leaders don’t want to part with their best people as they build their fiefdom! Nevertheless, the only way to retain and fully utilize your fastest growing high potential talent is to constantly position them in bigger and higher potential roles. Much like “re-potting” a fast growing tree allows the roots to go deeper and the tree to grow larger, people must also be repositioned to have the greatest impact. Sometimes this means shifting a product that is conceived in one part of an organization to a different team that is better positioned to fund, prioritize, and execute it. Other times it is about moving high potential emerging leaders to more mission-critical areas. Regardless, such shifts may feel disruptive (leaders don’t enjoy “losing” their best talent), but are incredibly necessary.

• Be willing to be misunderstood and poorly characterized before results materialize. The most impactful decisions working in a large company usually delay near-term progress in exchange for long-term benefits. Even though “short-term pain for long-term gain” is obvious, it is extremely difficult to justify any steps backward in a large public company with a quarterly cadence. The bigger the company, the shorter the attention span. You must be willing to be constantly questioned and misunderstood whenever you make long-term decisions with interim consequences. Doing a bold and entrepreneurial thing in a big company requires the tough skin to keep trying, despite missed deadlines from overly aggressive timelines (optimistic guesses that were necessary to get people to commit in the first place!). For major transformative projects (think major platform rewrites, restarting a product, etc) - the ones that most big companies rarely attempt and teams struggle to complete - they rarely ship on time, but need to be done anyway. Sometimes you need to just commit to the project and work tirelessly to do it with blinders on. You will need to tolerate soundbites about missed deadlines and periodic “why are we doing this!?” conversations, but everyone will enjoy the benefits of a major leap forward (and that’s all that really matters). The projects pursued despite great uncertainty yield the greatest outcomes. I was struck by the graphic below - the problem in big companies with short attention spans is that those zeros make people question the actions incessantly until they see the results. Defending the period of “relentless action before results” is half the battle.

• Promote unpopular people: Some of the most important people involved in the changes we made during my time at Adobe were also some of the most difficult to get promoted. Why? Because they were controversial. In some cases, they had made mistakes in their career because of the risks they were willing to take. During promotion discussions, people relished citing past shortcomings, as opposed to celebrating the bold bets taken and lessons learned. Often these leaders were ahead of their time, trying something just before the market was ready and then being punished for doing so. In other instances, people didn’t have the tenure we often use as a guidepost to making promotions, despite having a far greater impact than people who did. Making the case for these people, when there was always an easy reason to be a detractor in their promotion, was something I took seriously.

• Celebrate impact not tenure. One of the things I sought to eliminate, or at least de-emphasize, was the celebration of people based on how many years they had been at the company. Celebrating tenure sent a clear message to new talent that we celebrated longevity and loyalty over impact. As a leader, you must celebrate the people you want people to be like. Celebrate the people that are shipping, taking risks, clearing nonsense, and actually delighting customers. These are the people you want everyone to admire, rather than the people who have lasted the longest.

• Fight the soundbites. In a large and dynamic company there are so many decisions to make, and so little time to consider them. As a result, executives need to make snap judgments constantly - about people, about the status of projects, about the success of products. These “thin slice judgments” are often the only way to manage and stay afloat. But how do you make sure you’re not making horrendous decisions? You must obsess over data and adhere to a set of principles. On the data side, ground yourself in the facts rather than surrendering yourself to the soundbite. When people say something “didn’t work,” poke at the metrics. What time period is being used for this judgment? Are the KPIs the right objectives? Alongside the data, adhere to your principles. I wanted people to take risks and wanted to reward people who tried new ideas, even if those new ideas didn’t pan out. I wanted to encourage disagreement. I wanted to improve the brand and long-term customer relationships at the risk of near-term revenue. These principles helped me value people and projects despite the soundbites.

Note: If you share any snippets/screenshots with your own take on social, include @scottbelsky so I can share further/engage.

Ideas, missives & mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and my colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including some fun new tech I am playing with, my latest perspective on stories and commoditization of content, and several data provocations I continue to think about). Subscriptions go toward organizations I support including the Museum of Modern Art and fellowships for aspiring entrepreneurs at Cornell who will need some backing to start a company over the summer rather than take a typical internship. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) the much needed role of “safety layers” in our everyday lives, (2) how religion may complement the “all-knowing” nature of AI, (3) a critical truth behind “founders” and honoring one that was lost, and (4) some surprises at the end (including latest angel thesis!), as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • As LLMs remember memorize our conversations, preferences, and purchases, the value of sharing selective access to this memory with others will be the equivalent of a mind-meld. Here we discuss the implications of “Me becomes we” which has generated a lot of discussion since publication.

  • In April, we forecasted a new chapter in the ongoing battle to be the top interface layer, enabled by new tactics and renewed focus on having the best data underneath. And then last week we saw Slack start to cut off data APIs to startups like Glean. As this heats up, here are potential implications of the coming data wars.

  • Are we being reprogrammed by algorithms? Most definitely, and in this past edition we discussed the implications of our factory default primal human tendencies being overridden (the good and the bad).

• Alright, lets jump into Edition #32 of Implications:

The development of AI as the ultimate safety layer.

Being targeted in some sort of nefarious digital crime is becoming a daily occurrence. Scam calls and texts, phishing emails, sophisticated hacks disguised as docusigns, colleagues with cloned voices calling for wire transfers, and the list goes on. While people with less experience of tech are especially vulnerable, all of us are facing these tricks and puzzles these days and the precision and velocity of these crimes will only grow. I have nearly fallen for a few of these attempts myself over the years. There is much focus on how AI will make things worse (fake media, cloned voices, extended context windows to socialize us for a long-game scam, etc), but how might AI ultimately save us?

• Operating system-based AI safety layers to protect us. Whether we receive a scam via call, text or email, imagine a world where a “Safety Agent,” operating locally (on the device, to ensure privacy), is always scanning the data packets looking for evidence of digital tampering? Such an AI safety layer should have the ability to trigger a new type of OS-level warning or notification that is both audible and visible with a message along the lines of “This correspondence has a high likelihood of being a scam. Proceed with caution.” While phones are starting to do this based on certain phone numbers, future safety layers will be able to examine the content, the voice, the links, and any degree of socialization that may be fraudulent.

• Safety layers can also screen for behaviors with implications for mental health. If you’re doomscrolling past your ordinary bedtime, or you’re exhibiting patterns of depression as you engage in social media, the AI-powered safety layer could give you a similar message that “snaps you out of it,” if only on a temporary basis. Especially for young kids and teenagers, this could be super useful and entirely possible with an OS-level safety layer. I also imagine that AI will be able to diagnose the biases of the algorithms that govern our media consumption and notify us when we’re being polarized.

• AI-influenced judgment. We all know that texting or emailing while angry or anxious is often leads to our sentiment being misunderstood — causing issues between partners, friends, and colleagues. Perhaps the “safety layer” agents of the future can have settings that ultimately detect such moments and improve your judgement in real-time on how and when to communicate.

• Enterprise “always on” compliance: There is an opportunity for new AI applications to emerge in the enterprise, either as browser extensions or OS-level capabilities, that actually liberate employees to be more proactive and take more risk. These omnipresent Compliance agents would calculate cost/benefit in real-time and give people confidence to do things that would have otherwise been prohibited or stalled by some compliance process. AI should ultimately encourage agency and risk-taking while providing the equivalent of a real-time compliance agent that keeps a brand safe while eliminating cycles of review and approval.

• Can our AI safety layers have some empathy for us? Ok, now we’re getting a bit more “out there,” but AI sees more than what we end up saying or sending, it watches us struggle to articulate what we’re trying to say. Right now, as I type and then delete and then type again, my computer sees this internal monologue that you (the reader) never has. Perhaps AI will get to know us and help protect us in ways we cannot yet imagine?

• How do we balance the “big brother” nature of safety layers with their benefits? It is somewhat disconcerting to think about an AI, albeit one that runs locally on our devices, to see everything we do or view. But this cognitive barrier is more psychological than anything else. After all, our device already has all of this information on it. It just can’t be accessed and leveraged in these new and powerful ways. But I do think the ultimate solution here is a series of AI models that run locally, always updated to detect the latest threats while never transmitting our personal information to the cloud.

What is the role of religion as we navigate a world we increasingly don’t understand?

While the history of humankind is not particularly peaceful, the collective mysteries and miseries of humanity have never been reflected back to us. The notion of “God” in almost all religions is a one-way transom, through which we are welcome to pray and send messages (or sacrifices in more ancient times), but we don’t get clear responses.

During a long run this weekend, I was thinking about the current state of the world.

Nations dropping bombs on each other.

Gangs refining increasingly lethal drugs that people knowingly consume to escape this small precious window of life we get.

Governments racing to build the capacity to extinguish all people, playing an existential game of chicken with human existence in the balance.

People grabbing and squandering wealth at a breakneck pace while most of our fellow citizens and humans live in poverty.

Online personalities spreading falsehoods that polarize people, or worse, to simply accrue pageviews and clicks.

It struck me that we’re living in a world we increasingly don’t understand (or never understood), and the decisions we make as humans are increasingly against the interests of humanity. While religion is something people have turned to throughout history for some sense of morality and faith, the old doctrines also feel increasingly outdated. Regular church attendance has gone from ~49% in the 1950s to ~28% last year, and membership with any religious organization has fallen over the decades. In a modern world where every piece of information is at our fingertips, have people given up on the “one-way transom?”

Amidst all this, the advent of AI “super-intelligence” (many terms and definitions, but all researchers agree it’s coming fast) will be the first moment when something “all-knowing” (God-like) can express its opinion. As more of an Animist or Pantheist (awed by the power of nature) -- if not an atheist – I am surprised to find myself writing about religion in Implications. But one commonality across all religions is the depiction of God as “all-knowing,” omnipresent, etc…and I can’t help but wonder if one persistent feature (or bug?) of God is that it can’t respond nor say what it thinks of us. We cannot actually be judged nor clued into the mysteries while alive.

Fast forward to today, people can ask and discover what ChatGPT thinks of them based on all of their inquiries and interactions. Humans will be increasingly interested in what their AI agents and companions think of them as context windows for memory grow and models improve. For the first time, we’ll get a response from something “all-knowing” about us. The transom with ASI (artificial super-intelligence) will work both ways. As AI works in ways that exceed our comprehension and becomes more mysterious, will its perspective carry weight? Will its advice become a self-fulfilling prophesy? Will the next great religion-like relationship be one that is personalized to each of us? Before you scoff, consider that perhaps tens of thousands of people now view ChatGPT as a God.

The truth behind every founder is a founding team.

While I am proud to be a founder, I’ve always had issues with the official “founder” or “co-founder” titles because, in all truth, there is a “founding team” behind any successful business. In most cases, the founders of every function of a new company take an incredible amount of risk in their careers to build something from zero to one. They may not have come up with the original idea, signed the first lease, registered the domain name, or raised the initial seed capital, but they were critical ingredients and possessed amazing boldness, vision, ambition, and tolerance for risk.

In my experience starting Behance back in 2007, our founding team was the difference between this venture being a success or a failure. What is now a thriving network of creative professionals approaching 60 million members – a platform that has impacted countless careers and connected artists across genres in countless ways – started as a wayward venture with extremely inexperienced people building something we were not qualified to build.

We were lost for quite some time until a few key leaders joined us, including one particular ambitious, low-ego engineer named Bryan Latten. Bryan came from a family that valued hard work and never took a promising career for granted and he had a steady job at Lockheed Martin. So it must have been hard for him to make the leap and join our passionate yet clueless crew. But Bryan took the risk and jumped in, overhauling the architecture of our stack, refactoring critical parts of code, and instituting best practices in how we built and operated. He was very much a founding member of our team.

Most importantly, Bryan became a much-needed confidante and critic — always direct, passionate about quality and dispassionate about where the ideas came from. As our engineering team scaled, Bryan demonstrated the ability to manage others. If he hadn’t led us through these early years, I don’t think we would’ve found our way. Bryan epitomized the fact that “founding team” is a more accurate notion than “founder.” He founded so much of what Behance became, bringing essential skills, setting our culture, and finding our way through challenge after challenge.

Bryan joined Adobe upon our acquisition in 2012 and was quickly promoted to become one of the company’s youngest Vice Presidents. He left several years later to co-found a startup with another early Behance leader, COO Will Allen. (While I was sad to see them go, I was proud to be one of their first seed checks). Their startup was ultimately acquired by another startup. And then, just eleven months ago, Bryan called me with the startling news that he had been diagnosed with brain cancer. It was an emotional call as Bryan explained how he was preparing for the two possible paths ahead with the same directness and grounded optimism that he brought to every challenge we shared in our Behance days. Bryan gave this fight everything he could, but we all lost Bryan over Memorial Day weekend. Two young daughters lost an incredible father, his wife lost the best possible partner, his parents, sisters, family and friends lost a loyal and remarkable human being, and my former colleagues at Behance lost one of our founding members who made it all possible.

Over a dozen of us from the early days of Behance gathered at Bryan’s funeral to pay our respects. The common theme from our conversations was gratitude for Bryan’s friendship, leadership, and for saving all of our asses countless times over the course of our journey together. We all know that our journey and the success we shared, which has provided security for our families and opportunity in our careers, would not have been possible without Bryan.

The lesson? Build your founding team with precision and care and see them for who they are: founders. Perhaps “founding team” can become a more common part of the vernacular of entrepreneurship. Now, as I find myself starting A24 Labs and building an engineering, product, and design team, albeit within a larger organization, I have newfound appreciation for the founding team I have assembled. Each of us has taken a risk in our careers and is driven by ambition and vision for something that does not yet exist. Each of us believes that something new is needed in the industry of storytelling. And each of us feels in service to filmmakers and the broader teams that are required to bring world class stories to life. Here’s to Bryan, and to all the founding teams out there taking the leap and doing the work to push the world forward.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including my latest thesis points as an angel investor, some thoughts on what humans CANNOT be trained to do (as a proxy for what will always be out of reach of AI), and a series of data provocations that got me thinking... Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) our individual memories becoming part of a shared collective memory, (2) knowledge arbitrage during platform shifts, (3) the vastly underestimated impact of AI on education, (4) why the future of creativity is still bright, yet different, and (5) some surprises and off-the-record insights at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis (knowledge synthesis exercise) is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • AI-optimized algorithms are persuading us not with words, but with carefully choreographed feeds of content that slowly shift our opinions and desires. This may seem alarmist, but this edition explores the many positives (effective and inexpensive therapy, smoking cessation, tailor-made education) as well as potential negative (the override of human value systems, widespread shifts in opinions and values, etc) consequences of tech-reliance over self-reliance.

  • A platform shift causes us to reimagine everything while (and perhaps because) we become frustrated with the current soon-to-be-antiquated aspects of our lives that linger. Don’t miss these insights for reimagining your product during platform shifts.

  • Hyper-personalized pricing, on-the-fly UI and text-based-commerce, scarcity-driven offerings, AI that will haggle for you, and a return to non-scalable one-on-one attention from fellow humans in hospitality-driven experiences, here’s a set of wild expectations for the future of commerce.

  • Alright, let’s jump into Edition #31 of Implications…

Me Becomes We: The Implications of Collective Memory

As I watch new workflows emerge within AI-native companies, I am struck by how quickly everything done, found, or saved by individuals is enriching the collective memory of the organization. Products like Glean, Notion or Atlassian’s Rovo allow enterprise search across all forms of documents and data and products like Granola capture meeting notes. The default settings of these tools increasingly contribute our individual work to train the organization’s knowledge, which is available to everyone.

Consumer-facing technologies like ChatGPT and Claude are extending their memory of our conversations and it is only a matter of time before we’re given the option to share the AI’s memory of us with our family and loved ones — much like you share a photo album or a Spotify playlist. Fast forward, as LLMs remember our schedules, memorialize our conversations, and memorize our preferences and purchases, the value of sharing selective access to this memory with others will be the equivalent of a mind-meld. Your loved ones will be able to leverage — and even inherit — your accumulated knowledge. Their AI query for good restaurants in NYC or health concerns will yield responses that are enriched with your history and experiences. And at work, colleagues will be able to mine each other’s interactions and realizations, and network in unimaginable ways to advance a business.

This starts as shared memory (in the form of syncing data or selecting context windows (memories, organized by theme)) across a family or team. But eventually, this leads to literal brain melds when the Neuralink-like chips planted in our brains are enriched by real-time AI. I’ll be able to query my memory, the world’s knowledge, and shared memory with others.

In such a world, does “me” becomes “we”? What is it like to access parts of the accumulated knowledge and experiences of others just as you recall your own lessons learned? Who do you add to the “Disallow List” for access to your personal knowledge? Will your spouse and children expect to be a part of the collective memory of your family, or avoid it? In the enterprise, a new strata of status will be those who have access to collective memory and those who don’t. When we leave a job, does the enterprise get to retain and continued to leverage our memory? Also, in a world of shared memory, trust will be exceptionally important in the process of hiring new people or starting new relationships. When you hire someone — or marry someone — they will share parts of your context window with AI (your modern memory), and the implications cannot be understated.

Knowledge Arbitrage During Platform Shifts

I remember quite vividly in the 2007-2010 timeframe when very young “social native” experts (basically people in their early twenties who were deeply familiar with every social media platform and how to use each of them) capitalized on the fact that no large company or marketing executive understood these products or how to exploit them for their business. Friends of mine like Josh Spear and Gary Vaynerchuk, among others, launched their careers by becoming essential partners for some of the biggest brands in the world during this window of “knowledge arbitrage,” in which young, less experienced yet tech-native talent ran circles around functional experts across industries.

Well, here we are again. The myriad of creative uses of LLMs, prompt engineering, and the AI-ification of daily life are very familiar to young people in their twenties and a small cohort of early adopters, while completely foreign – if not frightening – to leaders of functions in large corporations. If you’re in high school or college and use ChatGPT and the long tail of generative AI models to hack your daily life in all sorts of creative ways, you have an amazing opportunity. The majority of others are just discovering all the utility and tricks that you now use reflexively. I encourage you to capitalize on this moment of knowledge arbitrage to launch your career while helping companies and leaders that desperately need your help.

Education is the ultimate low-hanging fruit of AI.

Fine-tuned AI will substantially elevate education for the average human. Obviously. I feel trite even making this assertion in Implications because, as you spend time with state-of-the-art LLMs and consider the impact of larger context windows and free tailor-made teaching experiences for every student on the planet, the impact is obvious.

For starters, 1:1 tutoring has always been the gold standard of education. Nothing beats a deeply knowledgeable and patient tutor with the ability to personalize a lesson plan to a student’s interests and learning style, and either speed up or slow down to optimize comprehension. Elad Gil shared this graph on the “two sigma improvement in learning achievement” of 1:1 tutoring (from research by Benjamin S. Bloom), which helps validate the advantage of tutoring.

Recently I’ve been trying to understand the key pillars of effective tutoring and mapping these pillars to the ideal design of AI agents…and the longer-term and perhaps less expected implications.

• Tutor for life, tailor-made for you: The more you use next generation AI tutors, the more tailored they will become. With an extended context window of every aspect of your learning style, struggles, interests, and strengths, the tutor will be optimized specifically for you in ways no human could ever be. Our tutors will also have the world’s data set at its disposal to triangulate the type of learner you are. We’ll look back and realize how ridiculous it was that traditional education failed to accommodate different learning styles, different interests, and forced us all to conform to a common curriculum.

• Variable pacing and personalized context: There are infinite ways to teach a concept, and pacing is never a constant. The ideal tutor can speed up whenever you get it, and slow down when you don’t. If a child likes horses or race cars, every worksheet and math problem can be reimagined for better engagement. MagicSchool is one company that is outfitting teachers with AI for this exact purpose.

• Market expansion of education: We typically think of education in traditional pillars, K-12, high school, college, and then on-the-job training modules. And then, in retirement, auditing a class becomes a luxury experience for those of us lucky enough to do it. But with the wide availability and accessibility of a hyper-personalized super-intelligent tutor at our fingertips, will education permeate our lives in unexpected ways? Will young kids with a passion for aviation proactively start tutorials on aerodynamics and go deep in advanced units of mastery out of sequence with traditional paths of education? Will our personal AI tutors proactively determine our interests and just take us down meandering paths of intrigue to discover and develop passions at our own pace that may greatly transcend and eventually antiquate the very concept of “grade levels”? And will adults develop new interests and masteries much later in our lives, enabling lateral moves across industries that were once unheard of (perhaps we’ll call these “vibe careers” in modern parlance!)?

The greatest impact of AI in education is confidence. When software meets you where you are, becomes tailor-made for you, and taps into the unique things that make each of us special, we will become so much more capable and confident in our everyday work and life.

The Future of Creativity is Bright, but Different

During my time leading creative and emerging products at Adobe, I spent a lot of time chronicling the workflows of many types of creators. And now, at A24, I spend time with filmmakers and other creators involved with storytelling, trying to understand how their process has changed as a result of all this new technology. Below is a graph I shared publicly during my last public keynote at Adobe, in an attempt to summarize some of the changes in the world of creativity.

Now, as I have shifted my time from building tools for creators to working side-by-side with them, the “new era” rings true. Modern tools that unlock 100x more cycles of ideation and concepting are making a profound impact on the creators who use them. Among writers, this might involve a deep conversation with an LLM to debate a character’s mindset in a particular scene or a request for a dozen scenarios for what might happen next to spark creativity. Among early adopting filmmakers, I have seen an explosion of time spent pre-visualizing characters, costumes, and particular scenes in films (especially on more fantastical or animated projects). With so many more cycles of origination and core creation unlocked by these new technologies, creators are able to explore more surface area of possibility in record time and ultimately make better choices.

In the world of marketing, I have seen creative teams shift their time from production (translating and reformatting endless assets) and refinement/quality checks (exhaustive proof-reading and checking copy) to exploring more ideas, going further afield, and allocating more time and energy to the final mile of polish to make things perfect … turning good into great.

The most exciting implication of all – and a major source of motivation for my small yet passionate team within A24 Labs – is to help the world’s greatest creators take more creative risk. The truth is, major creative projects like Hollywood films and breakthrough art are expensive. As a result, big studios shy away from risky (aka bold and new) stories and creators in favor of less risky ventures (like major franchises, known talent, and endless sequels). However, if new tools and techniques can help creators visualize, experiment, and explore the full landscape of their imagination with far less cost, perhaps we can advance new ideas at a fraction of the cost. Perhaps we can encourage and support more creative risk in ways that transform the world of storytelling and the lives of those who need these new stories the most. That’s my hope anyway.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor/product leader (including the renewed importance of science fiction in a world of exponential advancement, the unexpected role of writing, and data provocations on recent grads and federal land!). Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores a collection of insights and implications for your consideration, including (1) what product leaders must do to reimagine their products amidst a platform shift and what key learnings have revealed themselves, (2) what may transpire from the impending data wars, and (3) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • We talk about “agents” a lot these days, but what stages of functionality should we expect in the coming months and years across the products that embrace them? We explored this further here.

    

  • Amidst the rapid pace of change in technology these days, I recommend revisiting “the law of displacement speed,” which explains why, when applications or services are able to displace each other at a rapid and regular cadence, the result is either commoditization or platform-level displacement.

  • Finally, as the season warms for us east coasters, my daily runs have become far more enjoyable. Resurfacing 12 insights I hold dear, learned while running, which have had a big impact on how I lead and make decisions across various parts of my life.

And now, onto edition 30 of IMPLICATIONS.

Amidst a platform shift, what must product leaders do to reimagine their products?

The companies I am involved with, and the founders I advise, span the spectrum of products built pre and post the age of AI. Also, as part of my role at A24, I am constantly evaluating the products and processes behind every function of a fast-growing studio (a decent proxy for what many other high-growth businesses are now going through). I am struck by the speed at which pre-AI tools are feeling antiquated.

I recall the period when most New York toll booths went from being staffed by toll takers to “fast lanes” enabled by sensors, to the entire “toll” concept being ripped out altogether, replaced by high-speed cameras capturing your license as you speed by. Suddenly, the idea of having a toll-induced pause that causes any form of slowdown became unimaginable. My kids won’t even know what a tool booth is, let alone believe they even existed.

During platform shifts in which a new technology becomes widely available – like the transitions to electricity, to computers, to the internet, to mobile, to AI - you know the shift is real when tools and practices that we’ve used for many years suddenly start to feel like they are holding us back. A platform shift causes us to reimagine everything while (and perhaps because) we become frustrated with the current soon-to-be-antiquated aspects of our lives that linger.

I am feeling this sensation often. When I go into spreadsheets trying to make sense of various pieces of data, I realize that the lack of a fine-tuned LLM layer makes this an antiquated exercise. Spreadsheets are starting to feel like data silos compared to AI-native database products. When I use voice dictation on my iPhone or search for something and fail to find it due to an inexact descriptor, I notice the lack of AI and feel the cruft of rapidly antiquating software. The rules and filters that I set up in my native email clients are far inferior to the basic email agents that are emerging from startups.

Anything optimized for web or mobile rather than AI will start to feel outdated. How product teams react to this shift is crucial. Too many established companies will fall short because they only augment their offerings instead of fully reimagining them. So how should product leaders proceed, based on these implications?

• Re-anchor yourself with the customer. When you’re not sure what to build or how to build it, double down on empathy with the customer suffering the problem you want to solve. You’ll often learn that the solution you have in mind is a little (or a lot) off from what’s required to really engage the customer emotionally and practically. A platform shift challenges us to recalibrate our solutions. Are you making customers take manual steps that should now be automated? Should your defaults change completely based on what you now know about your customer? Amidst platform shifts, you need to start fresh, not just continue iterating.

• Go back to the first principles of human behavior and ask, “what do people naturally want?” The ultimate hint for the greatest product breakthroughs is whatever humans did hundreds of years ago, prior to the industrial revolution and technology as we know it. These are the things that we long to do again. Why? Because we are animals with deeply embedded natural primal tendencies from hundreds of thousands of years of evolution. Rather than fight our primal defaults, great product leaders leverage them. What was, and still is, natural? Things like reacting to what we see (recommended discovery and tailor-made user experiences will become more important). Talking to get answers (LLM-driven voice interfaces bring this back). Trusting recommendations from people we know more than average ratings from strangers (how will AI help us tap our social graphs, and how can products make recommendations feel more trusted?). Wanting to be known and feel special (we all have an innate desire to be recognized and remembered – how can the rise of agents with growing context windows for memory make us feel special, on our own terms?). The best technology takes us back to the way things once were, but with more scale and efficiency.

• Launch a new layer. If you find yourself managing products that struggle with utilization of the incremental AI features you are adding, take a step back and ask yourself, “What might an entirely new layer or overlay feel like?” With AI, you have the opportunity to use machine learning to leverage all the data and workflows you’ve developed over years via an entirely new layer on top of it all. One great example I’ve been watching in real-time, as a board member and customer, is Atlassian launching “Rovo,” their AI search and agent framework that sits on top of all Atlassian products. Rovo connects not only to Jira, Confluence, Loom and other Atlassian apps but also features more than 50 connectors to data that lives in Slack, Airtable, and many other sources. Another example is emerging products, like Softr, that are dedicated app builders agnostic to the sources of underlying data. A common piece of advice I give founders and product leaders these days: add a new AI-driven layer to your product before someone else does. Don’t relegate yourself to becoming a data repository for someone else’s interface!

• Indoctrinate your learnings. It’s easy to get caught up in the “breakthrough of the day” and churn product roadmaps too easily as a result. Like many of you, I’m trying to indoctrinate the learnings from others as I seek to build and advise on the product front during a platform shift. Here’s a quick hit list of learnings that have stuck with me, some more obvious than others.

  • If you’re building foundational models, you better move up the stack fast. The Law of Displacement Speed is proving itself to be true as models become further commoditized and indistinguishable.

  • You must capture and leverage first-party data to drive personalization.

  • You must build connectors for a very large number of other apps that your customer uses to further personalize your product experience with third-party data. But this won’t be differentiating for long — and will be harder after the impending bilateral data wars discussed in the next section.

  • You must capture some sort of social or professional graph to make recommendations and algorithms more relevant for the data you have.

  • If your system prompts are your moat, your days are numbered. These are being shared or uncovered at high velocity.

  • Your AI’s context window must be increasingly wide and long, since “personalization effects” are the new network effects when it comes to retention and product-led-growth.

  • Markets will be won or lost at the interface layer, and the ultimate interfaces are ULTIMATELY controlled by the operating systems of our lives (more on this below).

The Data Wars & Interface Layers

Of course, “don’t let your products become a data repository for someone else’s interface” is defensive advice. The more offensive version of this is to “build a killer interface and hyper-personalized and intelligent AI offering leveraging all the data you can get - and hopefully some data nobody else can get.” As I scan the ecosystem of successful AI-native products across every category of software, a common theme is data syncing and connectors. Glean, Atlassian, and a number of other players in the enterprise search space are rapidly creating data connectors enabling teams to sync all of their sources of data so they can search and manage all permissions centrally. New AI-native (anti-traditional-CRM) solutions like Folk and Clay sync with your GMail and social accounts to leverage all of this data to unlock new forms of contact search and discovery. New AI-native workflow apps like Scribe, and end-to-end enterprise-grade AI platforms like Writer (both investments/founders I support), are capturing every day workflows that people in companies do, remembering and training based on this data, and then automating entire workflows within functions of an enterprise. And the list goes on. Smart teams are building new interface layers (great opportunity for UX ingenuity!) that are enriched by data. The winning products will have their own first-party (hopefully exclusive) sources of data, combined with every other necessary data source, brought to life using proprietary methods for personalization and relevance. And all this is leading to the impending “data wars”

We are at the precipice of a new chapter in the ongoing battle to be the top interface layer, enabled by new tactics and renewed focus on having the best data underneath. For consumers, the battle will be fought primarily across operating systems – we will each have a default mobile OS, desktop OS, browser, AI-powered chat and search tool, and an immersive OS. In work tools, the battle will be across every function in the enterprise. On the consumer side, today’s reigning operating systems are racing to catch up , while new approaches like Meta’s AR explorations emerge. On the enterprise side, dozens of big companies and startups alike are building data connectors for the long tail of apps that hold data critical for AI to connect dots in search and ultimately perform work with agentic capabilities. As this data becomes more valuable and coveted, what might we expect to see?

• Bilateral agreements and game theory will be a hot topic among executive teams as they decide whether to cooperate with companies building data connectors and when (and how) to shut down or restrict data access.

• Data scraping and training via vision models will become a default. New tactics will emerge to gather data without requiring connectors, like employing client-side, multi-modal models with vision that leverage real-time screen monitoring to capture the necessary data and workflows. I also like how Scribe, mentioned above, lets any member of a team click “record” at any moment — across any application they are in — to capture a workflow of any kind, and then share it with all of their colleagues (and save it for their own future reference as well). If you take a step back from this, you see the genius in this being (1) a product interface layer that can sit and operate above all others and (2) a training data source for the workflows across an organization. What started as a simple browser extension has become an incredibly successful company still under the radar of most tech VCs and leaders (except those reading right now!).

• BYOD (bring your own data) tactics: Product makers will embrace imaginative efforts to let you bring your own data, perhaps mounting underlying repositories of data (think gaining direct access to the Snowflake or Databricks instance used by a third-party tool you subscribe to).

• Permission management and data security will become one of the most distinguishing features for enterprise tools in the age of AI. This is extremely hard work across engineering, architecture, and design that very few companies get right.

• Recommendations and signals based on interpersonal behavior become key differentiators. Social and work graphs will offer algorithmic advantages to those who understand who is connected to whom and who works with whom.

• Personalization effects are the new network effects. Context windows and memory quality will become a differentiator as well as the primary retention and “product led growth” tactic. The more you are remembered, the more tailor-made the product will become for you, the more likely you are to share with colleagues, and the less likely you are to churn. A product’s personalization becomes the new engine for growth and retention.

I’ve said many times in previous editions of implications that the “winners” in the AI era will be differentiated through interfaces and data, not the AI models in between (only a few exceptions). I’ve also said that the majority of the value in the era of AI will accrue to those who use models as opposed to those who make them - much like what’s happened to utilities in the era of electricity. Now, as time passes, I believe the data wars will be the most important battleground for determining the winners and losers in the era ahead.

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including the “market expansion” of education, determining the gullibility of AI, a team out of Barcelona helping people develop synthetic memories, and a few other observations). Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) how humanity is being reprogrammed as our primal instincts are subconsciously overridden (prepare yourself for this one), (2) what a future “History of Tech” class would teach us about today, and (3) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • The rise of cognition-driven companies (aka “cognicos”) and what I’ve come to call “the cognition stack.”

  • We have an ongoing series of insights for the modern product leader, and part two discusses why the product leaders who win their industry often do so by delaying gratification, how you should get customers to talk about their problem rather than your product, and why the best product leaders are optimistic about the future yet pessimistic about the present.

  • A set of wild expectations for the future of commerce, including context-based purchase decisions, hyper-personalized pricing, on-the-fly UI and text-based-commerce, more scarcity-driven offerings, AI that will haggle for you, and a return to non-scalable one-on-one attention from fellow humans in hospitality-driven experiences.

The reprogramming of humanity as primal instincts become overridden by AI.

I often ask myself, in the process of writing Implications, what is the most important thing we aren’t talking about? A recurring theme is the consequences of tech-reliance over self-reliance. For example, we discussed how (and why) algorithms that are optimized for engagement are socializing us at scale. This is important because it is an example of humanity being programmed by technology, rather than the other way around. If you’ve ever participated or witnessed a debate with a state-of-the-art LLM that is directed to convince you of something, you’ll see how effective AI can be at persuasion. Now imagine that the AI is persuading you not with words, but with carefully choreographed feeds of content that slowly shift your opinions and desires. This may seem alarmist, but I see as many positives (effective and inexpensive therapy, smoking cessation, tailor-made education) as I do negatives (the override of human value systems, widespread shifts in opinions and values, etc). Let’s explore the implications of world-class AI influencing (if not programming) of our decisions, opinions, and belief systems.

• AI advice and predictions become self-fulfilling prophesies. We aren’t too far from trusting AI more than our own instincts, at least in certain parts of our life. Whether it is a commerce decision about what inflatable raft to buy, at driving route to take or what restaurant to visit, we will increasingly have faith that AI (as a compendium of the world’s information and reasoning) has more credibility than any other source. As AI becomes increasingly personalized and powerful, we will stop questioning it. Which begs a very important question: Will advice and predictions from these AI tools become self-fulfilling? I saw some people asking the latest DeepSeek LLM about Bitcoin price predictions and other future oriented predictions. The answers and the exposed reasoning layer that accompanied the predictions were extraordinary. This all made me wonder about whether, as we trust AI more than ourselves, the guidance and predictions from AI will become self-fulfilling prophesies? If AI tells us, based on everything it knows about our finances, family, actions, and values, that we should make a particular purchase decision, career move, or vote for a particular candidate, at what point does our trust become blind faith? Will there still be people who are willing to test the AI’s effectiveness by disregarding its advice? And, at scale, when does blind faith become self-fulfilling to the point where AI predictions are determining the outcomes?

• Are we being reprogrammed by algorithms? In the last edition, I mentioned a recent tweet-er-xeet post by Twitter founder Jack Dorsey. Jack suggested that we should all be able to choose the algorithm that controls our timelines across social media products. But the underlying premise is more interesting: As social algorithms get far more effective at commanding our attention and tuning our interests, they remain a mystery to end users. We certainly cannot understand how they influence us if we don’t understand how they work. However, if consumers had the option of choosing an algorithm much like we choose an app on the App Store, we would be forced to understand how these algorithms work and decide which ones we want. This decision-making process would expose the objectives of each algorithm, does it give us more of what we like and agree with (filter bubble algorithms) or challenge our views with opposing narratives? Some algorithms may be optimized for humor, content from friends and people we actively connect with, or news from both sides rather than reinforcing one particular side. Some algorithms could be customized to help change your behaviors for the better — like adding a bias towards improving you as a parent, improving your wellness lifestyle, or developing you culturally through exposure to great music, films, and original art.

• Is the next step in human evolution an “override” of our primal instincts? As humans, we have a set of “factory defaults” that have persisted for ~300,000 years. Whether it is sexual attraction and the desire to procreate, the instinct to ensure shelter and protect your children (through, for instance, a steady job), or the “fight or flight” mechanisms that channel our fears and prejudices into instinctual biases, there is an undeniable set of default drives that govern so many of our natural human tendencies. So much of modern society is designed to help suppress or channel some of our more destructive primal defaults, with mixed levels of success. But as I considered the accelerating pace of AI model innovation and Jack Dorsey’s proposal that we all choose our own algorithms, I wondered when (not if) these AI-powered algorithms become effective enough that they override our factory default settings. When might these algorithms of persuasive content and doses of personalized social proof influence our hunger (say, convincing us to use supplements instead of naturally satiating our appetites)? When might algorithms influence culture and society’s definition of “beauty” faster than biological factors that drive procreation (and evolution) via natural selection? When might other decisions we normally make out of primal instincts - from the leaders we vote for to the friends we keep and the lovers we choose - become governed more by AI-optimized algorithms than anything else?

• Optimizing for success vs. Optimizing for agency. Until now, we have all lived rather self-determined lives riddled with human error. But the era ahead offers the chance to have every decision we make take the world’s knowledge and probability analysis into consideration. In such a world, success is most often assured by doing what the AI tells us to do. But what we gain in eliminating human error, we lose in agency. Sure, we technically have a choice, but going against the guidance of AI will increasingly look reckless if not self-destructive — much like turning off the headlights while driving at night (driving manually, that is! damnit my metaphors are losing steam in the age of AI!). As humans, we will need to decide where classic self-determination adds value through originality and creativity, and where it is simply a form of self-sabotage. How many people will choose to live partially or completely AI-free, like the people who choose to live off the grid now, and how tolerant will society at large be of that choice?

• We are being programmed. There’s no delicate way to say it. The technology to reprogram ourselves — and unduly influence the desires and beliefs of others — has arrived. Our daily life is flush with evidence of the power of AI-powered algorithms, from the radical and volatile shifts in political views and “cancel culture” to the speed of market trends and spread of memes. The water boils slowly, but we are the frogs without a doubt. How do we respond? I have three things to say: (1) we must take control of the programming by understanding how algorithms work, demanding transparency into how they are optimized, and exercising our choice for how we want to be influenced; (2) let’s use AI to program ourselves in the same way we take courses to learn and change our behaviors in our professional and personal lives; (3) let’s reimagine the full stack of education (version 1.0 of human programming was good ‘ol classroom teaching) using this new tailor-made approach to programming the next generation to be skilled, creative, passionate, and values-driven leaders and thinkers. Only by taking the reins on this technology can we ensure that this is, in fact, an upgrade of humanity as we know it.

What would a future “History of Tech” class teach us?

Sometimes we need to do post-mortems on our industry, not just our own projects. Especially as I review new startups, think about the future system of work as an Atlassian board member, and build out A24 Labs to provide the world’s greatest storytellers with the absolute best stack of technology, I have been spending time synthesizing observations and lessons learned. The fun prompt here is, “What would a future “History of Tech” class teach us?” I have a growing collection of these conclusions, but here are a few…

1. Competition between technologies is ultimately a game of “slap a hand” in which the layer on the top supplants the technologies beneath it…until a new layer is built on top. Companies made apps, and then OS-level features supplanted many of those apps. The default search in your browser supplanted the sites you went to for search. Chat-based agents are now supplanting the discovery process and destinations we once frequented to find answers (soon we will ask for a car, as opposed to going to a particular branded app to order one). Enterprise search and function-specific agents will gradually supplant the places you visit (and the people you go to) to get data and analysis. If you want to disrupt any industry — or prevent your disruption — go up the stack of user experience.

2. Novelty precedes utility. I remember when Slack first came on the scene over a decade ago, when my team at Behance was using HipChat. We had no need for another messaging app, but people loved using Slack to summon company when they were leaving the office to grab a coffee or leverage “/giphy” to send random animated GIFs to each other with some degree of plausible deniability. What started as fun led to the discovery of utility and ultimately Slack became a mission-critical technology for the team. I saw the same phenomenon during the rollout of a virtual conference room technology in 2005 while working at an investment bank. Nobody used it until one member of the team summoned everyone to “jump on audio conf room 1” to secretly make fun of a partner’s tie. When we play with new technology, we become socialized to its use cases. Now, in the age of AI, the same pattern is repeating itself and we must let our teams play to discover the utility.

3. Don’t dismiss or discount new tech because of early misuses. Whether it was the early use cases of the internet for porn and impropriety, the early use cases of Bitcoin for illicit transactions, the early use cases of NFTs for scams, the early use of Generative AI to make deep fakes, and the list goes on…we must learn to look beyond the early uses cases of new technology, and we must wait to pass judgment.

4. Technology ultimately succeeds because of the user’s experience of the technology, not the technology itself. Fighting words here, and I don’t mean to offend my engineering colleagues, but have seen this prove itself time and time again. Breakthrough technology spreads fast and is seldom the moat you expect it to be. Look no further than the rapid commoditization of LLMs and media models, the ubiquity of mobile apps doing everything you can imagine, or the dozens of SaaS companies that emerge to solve problems faced across every function of business and life. The tech itself is important, but great tech alone does not guarantee success. The user experience determines whether a new customer can survive the first mile of the product, whether the product’s functionality is even used, whether the customer is willing to pay, and whether the product grows. The secret of any successful and honest product leader is their design partner. When you empower designers at every part of the process of building products — and companies — you stack the deck in your favor. What you’ll also learn is that design can compensate for technical shortcomings. I saw this in the early days of Behance, Pinterest, Uber and other companies that had growing pains in the form of technical scaling and performance issues that were addressed, at least initially, with design changes.

5. Momentum is a moat. As an optimistic entrepreneur, chief product officer focused on innovation, and especially as a leader of M&A for years, this is a lesson I have learned the hard way more times than I care to admit. When a product has escape velocity, it doesn’t even need to be the best product on the market to continue winning. This pains me to say, but history proves it time and time again. Why? Because the vast majority of potential customers for a product are pragmatists and only adopt AFTER rampant adoption. Most companies (and consumers) choose the safe option when it comes to tools. Also, the ripples of network effects continue far longer than people expect. You need to hear about a product many times before you’re willing to try it, and any product with a learning curve, once overcome, is far stickier than the average entrepreneur imagines. Most people take “don’t fix it if it ain’t broke” too literally, even if there’s a better way.

6. Insanity is a moat. There’s no better way to overcome an incumbent’s moat of momentum than using a dose of insanity. I’m being cheeky here, but Airbnb was “insane” enough (”who would ever let a stranger sleep in their home!?”) to prevent the entire hospitality industry from competing for over a decade. When you launch a new product or tool that is border line “sacrilegious” to conventions, you turn heads enough to attract great talent and leapfrog competition. We’ve seen this with Anduril (“”defense startups are impossible”), Rippling (”good luck competing with entrenched leaders and practices!”), and, Figma (”the web will never be reliable enough for professional design!”), among other examples that stomached being misunderstood long enough to change their industry.

7. Open source almost always exceeds expectations. The power of collective contribution and creativity outperforms centralized teams and bureaucracy, time and time again. So many companies are desperate for a business model out of the gate and fail to quantify the value of the world working to build and maintain their technology for free.

Building for The Future of Storytelling

Building for The Future of Storytelling: I mentioned in a previous edition of Implications that I have joined A24’s leadership team and am building a team of remarkable designers, engineers, operators, and technologists who all share an obsession with storytelling and empowering creative minds to take more risk and tell world class stories in new ways. Are you a designer or engineer who loves building products that stitch together various parts of the creative process? Are you a comfyUI expert who wants to push new media production technologies to their limits? Are you an operator who loves outfitting fast-growing companies with better tools to coordinate and execute ambitious projects? Or do you know someone I need to meet? Reach out and let me know!

Happy birthday Dot Grid Books! It was ~17 years ago this month when Matias and I designed and launched a few new additions to the Action Method product line, including the “Dot Grid Book.” While originally designed for ourselves, so many designers, architects, illustrators, and product designers started carrying these around. The Dot Grid Book was developed as an alternative to traditional lines and boxes, using a light geometric dot matrix as a subtle guide for your notations and sketches. It comes in regular and a smaller “Dot Grid Book Mini” size (my favorite). To commemorate, the team keeping this dream alive made a 20% discount code for Implications readers, valid until the end of the month. DOTGRIDBDAY20

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues past and present…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind as an investor, technologist, and product leader (including the rise of biologically-inspired computing and a new investment I have made, my generally negative outlook on venture capital as an asset class (and reasons why), and other random thoughts. Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) how friction was once, but will no longer be, a strategy, (2) the stages of functionality of AI agents, (3) the growing proliferation of gamification, (4) another round of insights for the modern product leader, and (5) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • Conformative software is a new wave of apps that become more tailor-made the more you use them. These apps are designed and built with a new form of accommodative user interfaces that adjust to your skill level, your needs, and your preferences.

  • A few things I expect to see in 2025, from new use cases of betting markets and DIY software to the rise of Cognico companies and AI talent flocking from the companies that build AI to the industries that will be most transformed by AI.

  • Moving past the prompt-based era of creativity, unleashing creative risk, and creative meritocracy.

But first…Behold the world of storytelling.

A couple weeks ago, I shared a shift in my own career as I transition from leading strategy, design, and emerging products at Adobe to joining A24, an independent studio I have long admired, as a partner and leader of new efforts across innovation and technology. I will miss my teams and products at Adobe, but also couldn’t be more proud of the leadership team we’ve built and the roadmap ahead. I am also excited to spend more time as a customer using some of the products and services we’ve made over the years (and especially some of what’s coming soon!).

A24 is a special team and place, and has always had a model that not only empowers talent but ultimately works in partnership with the world’s greatest storytellers (a friend sent me a good write-up on A24’s origin story here). Suffice to say, I’m excited for the steep learning curve ahead and applying all I’ve learned serving the creative industry through Behance, Adobe, and my side projects/writing towards A24’s business and future. Bigger picture, as the space of media and technology evolve and many “skills” are offloaded to compute, I believe that taste and curation will be key differentiators. In an era of abundance in which every brand floods the zone with content, I think humans will crave better stories, with deeper meaning, craft, and new forms and mediums of shared experiences. And amidst so much volatility around the world, I have long believed that stories, in the form of films, shows, books, and music, are the most effective methods to transport payloads of hard truths.

I hope for this monthly “implications” exercise to continue, because it - and you - have helped me synthesize so many changes in technology and culture over the years — and their implications. It turns out that a wide spectrum of inputs, in the form of ideas and experiences and conversations and feedback, impacts your outputs. For this I am grateful.

Assertions & Implications

Lets jump into three assertions with implications worth noting, as well as another section of insights for the modern product leader.

• Friction is no longer a strategy.

• Deconstructing “agents” into the stages of functionality.

• Gamification of everything.

• Insights from The Modern Product Leader Playbook, Part 4

Friction is no longer a strategy.

One early use case of AI’s new “Operator” capability that I found particularly intriguing was people tasking the AI to submit claims - whether to health insurance, for refunds, entering online sweepstakes, claiming redemptions of credits, getting payment from a class action lawsuit, or other means of holding companies accountable for their obligations - in repeated and increasingly creative ways to ultimately get payment. Whereas most humans who value their time will tire with these activities, AI never tires and has no fear of failure. As a result, as these “automated browser type” capabilities become more widely available and fine-tuned for particular use cases, companies can no longer rely on the additional margin accrued from the “friction” of these activities. This is wonderful for consumers, at least initially, but may ultimately result in new policies or higher prices to account for this new world.

One company that has been ahead of the pack on this is DoNotPay, a service that bills itself as “the AI consumer companion” and has tried to productize some of these capabilities - like contesting a parking ticket or getting refunded when your internet goes out. But I suspect many other tools and AI-powered hacks will emerge wherever human friction remains a strategy of sorts for companies and governments. The use cases of AI removing friction to foster equal opportunity and righteousness are seldom discussed but very interesting. The people who suffer the most from intentional friction are often the most disadvantaged (who also have the least time to spare). Let’s hope AI is a source of productive change.

Deconstructing “agents” into the stages of functionality.

This is a topic I am brewing for a future issue of Implications, but figured I’d share some early thoughts on the agent development efforts I am seeing across startups I meet with or within big enterprise projects, and apply some organization to these functionalities. Here are the five stages of agent capabilities as I see them at the moment:

• Glorified Personalized Help: The most basic (and least impressive) functionality of agents is glorified personalized help akin to an FAQ being more personalized for you and the problem you’re trying to solve.

• Reactive Recommendations: You request things and get responses with some degree of work done on your behalf, whether you are summoning code, assets, or some other prepared personalized useful artifact based on your request. Most startups pitching agent-based functionality, and the agents I have played with, fall into this camp.

• Proactive Recommendations: The next stage is when agents within a product experience or workflow proactively suggest something you may want to do but didn’t think of. Such agents often require multi-modal LLM models under the hood that can see what you’re doing, ascertain the context in which you are doing it and what you hope to accomplish, and can proactively and reliably make recommendations. Apart from structured systems like code generation workflows, I have not seen many great examples of this yet…but they are coming.

• Proactive Action: Going further up the pyramid of possibility, the fourth stage is when agents not only proactively recommend things we might want, but actually do these things for us. This is the stage where an agent becomes a collaborator or colleague — where we start working side-by-side. We’re seeing this form of agent emerge in some coding applications, and I am excited to see this vision really come to life.

• Autonomous Workflows: Finally, the upper most stage that I can imagine at the moment is fully autonomous agents that run end-to-end workflows on their own. Such workflows may involve agents making purchases on our behalf, agents working (or negotiating with) other agents on our behalf, and agents staffed with specific jobs and OKRs (objective key results) that they must meet. There are so many exciting implications of this stage, including new tools for observability and new human roles in organizations to set the goals and “rules” for agents to abide by. I explored the vast implications of this level of capability in my “Cognicos” edition.

Finally, a major caveat or rebuttal to my own outlook and analysis: The challenge to this way of thinking is whether or not we are “personifying” AI too much and trying to figure out how it fits with how humans work. When we talk about the rise of “virtual AI colleagues” are we falling into the “faster horse” trap that has always delayed the value realization of any new technology? I remember the early days of the iPhone, when Apple used skeumorphic design principles like adding fake physical thread stitching to the contacts app to make it look more leather bound and help us all adopt these new technologies through familiar patterns. But longer term, these efforts to make technology familiar ended up localizing the potential outcomes. How do we think about the use of “agents” without the biases and assumptions we have from the context of human work over the last 300,000 years? Perhaps this is a topic for another edition… ;-)

Gamification of everything: widespread and unexpected proliferation of game mechanics.

There aren’t many apps that would NOT benefit from gamification, whether it is in the form of leader boards, native currencies and rewards, daily challenges, etc. While gamification tactics has been used in popular consumer apps (Duolingo, “achievements” in diet and fitness applications, medals on our Apple Watches, etc) have been used for years, I have long wondered why such mechanisms wouldn’t work in the enterprise? Don’t we enjoy a meeting or work more when it’s fun? What if daily practices and accomplishments in the workplace had more gamified elements? Would such mechanisms increase our engagement and performance in school or at work, much like as they do in the games that we play?

As an angel investor, I am seeing more pitches in spaces beyond traditional consumer tech that incorporate game mechanics. Some education technology companies are developing entirely new game mechanics that make learning more exciting and competitive amongst peers. Good Inside, a company devoted to equipping parents with scripts, workshops, and guidance to navigate all sorts of parenthood challenges (where my wife is co-founder and COO) is also exploring the role of game mechanics to help parents navigate the many modules and milestones of parenting. Another team in the dev ops space I advise is exploring gamification of code commits and general efficiency and effectiveness for developer teams in the form of leaderboards and other tactics that will be cool to talk about once launched. (Wand when I saw recent prototypes, this felt super- powerful to me.).

This wave of adoption is not just about the readiness of technology or the tactics to drive gamification, it is also about culture. As one founder explained to me, “over 50% of kids under 12 in the U.S. play Minecraft, and 50% of kids under 16 play Roblox. These kids spend almost two hours playing each day and most of their lives are virtual and in-game, from attending concerts to hosting parties.” The logic is, if the next generation of talent is so immersed in game mechanics throughout their childhood, why not capitalize on these patterns and tendencies across other aspects of their digital lives when they become adults? Especially for those born after the emergence of the smartphone, gamification leverages a set of reflexes for engagement and self-optimization that run strong.

The challenge will be incorporating meaningful incentives while avoiding situations in which people “game” work. We already know that some people manipulate OKRs and other measures to make their work seem more impactful than it is. We’ll need to make sure that gamification is designed and implemented in such a way that gets us doing more of what we should be doing, and less of what we shouldn’t.

Insights from The Modern Product Leader Playbook, Part 4

Here is the fourth installment of a set of insights I share with or learn from modern product leaders I admire. Pretty sure I’m done writing books, so figured this was an optimal place and audience to share these insights.

• As a startup, you should only do half of what you want to do (only half the options, half the tabs, half the offerings, and half the target audience) to compound your chances of true PMF (product-market fit). Most founders I advise and work with are incredibly ambitious and optimistic — two traits that help and hurt when building a product that really nails its core value. Early on in my product-building days at Behance, I often wanted to hedge our team’s likelihood of success by including more features given I lacked the certainty of which features would ultimately work. But over time I learned that, when you kill secondary features, people engage with the primary purpose of the product more. The truly experienced product builders are like Bonsai masters, they prune and cut the most beautiful branches to strengthen the trunk.

• Natural > Rational. Great products capitalize on natural tendencies of humanity. Consider this thought experiment: You’re on a subway and overhear two people you don’t know talking about a downtown restaurant as the greatest Italian place they’ve ever been to. If you’re in the mood for Italian food, which opinion do you trust more, the two strangers talking on the train or the average rating of lots of strangers on Yelp? Rationally, you should trust Yelp’s average more. But most people would go with the overheard opinion on the subway. For some reason, people tend to trust people over averages and networks. Even though this isn’t completely rational, it feels natural. There are lots of these instances in which we choose what feels natural over what is strictly rational. Valuing natural tendencies over rational thinking can help you make your products more engaging and better able to hook customers faster than the time it takes for them to realize long-term value. The human brain hasn’t evolved so much in hundreds of thousands of years, so it is worth asking yourself, across each part of your user experience and especially at choice points, “what would the natural and familiar instinct of a user be?” When designing a product (and the copy, options, and UX), seek to understand your user’s natural tendencies rather than what they should rationally do.

• Copy what drives familiarity, not what sets your product apart. In early Behance days, we wanted to launch a create spin on hosting group conversations for creative pros online (we called them “Circles”) and having people identify their creative fields of interest (we called them “Realms”). And in the effort to be creative and different with these features, we made them harder to understand. Ultimately, we renamed these features Groups and Fields, and then eventually integrated them into other parts of the product (basically removed them all together). The classic mistake we made was failing to use familiar patterns every chance we got. Over the years working with dozens of other products and founders, it became very clear that you should make ~90% of your product using familiar patterns, and only retrain your users ~10% of the time for the few parts of your product that are truly differentiating. Ask yourself: What is your product do that is unique? It is these few specific things - and the functions that enable them - that you should neither outsource nor emulate. As for everything else, copy what works and don’t reinvent the wheel.

Ideas, Missives & Mentions

**Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, and colleagues…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and remained on my mind, including how algorithms are programming us, whether AI advice will become self-fulfilling and what’s happening in in-car advertising, among other areas of interest. Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) wild expectations for the future of commerce, (2) the era of “conformative software” that becomes more tailor-made as you use it, and (3) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • A few things I expect to see in 2025, from investigative journalists leveraging betting markets to top AI talent shifting from working in the companies pioneering AI to taking leadership roles in the industries that will benefit most from AI.

  • The topic of augmented reality (AR) right now feels a lot like AI felt ~2016 when everyone knew AI would be important for the mainstream but the building blocks just weren’t there yet. We’re getting closer…

  • The rise of “cognicos” - Until now, humans have been the reasoning layer of every organization, but a new era of cognition-driven companies will change companies as we know them.

Wild expectations for the future of commerce.

One part of daily life likely to feel entirely different within a few years is commerce. Our digital commerce experiences will be hyper-personalized; we’ll be presented with a selection that feels catered to our preferences and accounts for past purchases. Our AI-powered agents will make pre-approved payments and micro-payments on our behalf and will generally make day-to-day commerce more seamless than ever. New “on the fly UIs (user interfaces)” will simplify choices, catering to our needs while obfuscating (aka disrupting) the underlying brands and providers. And the AI on our devices will protect us from anything fraudulent, proactively warn us of scams, and assist us in finding good deals by evaluating customer feedback and consumer reports before every purchase. In short, commerce is going to feel very different.

• Context-based purchase decisions. Imagine every purchase decision - from food items and vitamins to wardrobe and accessories - being framed in the context of your diet, what you’ve purchased before, or what is recommended based on a deep analysis of your life and preferences. When you want to buy something, your personal AI will make the best recommendation and explain why. Of course, in the process you will begin to care less about the brand (and potentially be less persuaded by marketing) as you begin to trust your personalize AI (given its context and reasoning) more than any other signal.

• Pricing will become hyper-personalized based on our loyalty, preferences, and willingness to pay. Imagine special offers extended to customers based on, among other ideas, their taste, influence on social platforms, and their viral co-efficient (i.e. their willingness to share information about their purchase in ways that yields other customers, a measure that could certainly be determined from past purchases in the age of AI). AI-driven pricing is a fascinating and somewhat disconcerting phenomenon about to gain new dimensions and go mainstream.

• On-the-fly UI & text-based-commerce will increasingly grab a share of our daily commerce experiences. Imagine telling your OS-level AI companion - or some other AI Agent - that you need to buy protein shakes or want to book a trip to visit a friend in Boston. Instead of going to a marketplace like Amazon or a travel website, the AI experience can serve you a custom interface (like a mini ephemeral app made just for you in this moment) that helps you browse choices and make decisions. In doing so, this “on-the-fly UI” will present everything in the context of your preferences. You may not even see or care about the brands involved or who is actually fulfilling the product or service. The same thing will happen with the next generation of commercial texting/messaging. These experiences will shift from being spammy annoyances to being extraordinarily helpful once they become super-intelligent and actionable. While I happen to HATE text messages from stores or brands, what if texting with a brand is more akin to a personal shopper that has worked with you for years, remembers your preferences (whether your wardrobe or otherwise), and can really engage with you personally and act on your behalf?

• A return to non-scalable one-on-one attention from fellow humans in hospitality-driven experiences and the final mile of any purchase. I’ve been anticipating, in previous editions of Implications (like this one on “the meaning economy”), that human jobs would shift towards less scalable areas that are differentiating for brands, and support is surprisingly one of these areas. I am not talking about call centers, but I do love the idea of real people helping you make decisions or devoting their time to make your experience magical. I have long believed that the science of business is scaling, and the art of business is things that don’t scale. With the massive savings of automating customer support and many other functions of the enterprise with AI, there will be a drive to differentiate customer experience and brand with good old-fashioned relationship-driven human attention. There will be more time for the art of business.

• “AI will haggle for you, but smartly.” This was an insight from Greg Isenberg, who shared a collection of other thoughts about the future of commerce in his newsletter, a few of which I wanted to add to the list. In this case, AI informing multiple merchants of your intention to purchase something, and then allowing those merchants to bid for your business.

• More scarcity-driven merchandising in the age of abundance. Several friends have recently mentioned the growing phenomenon, especially in the high-end handbag industry, of no longer allowing real-time purchase of high-end goods and instead requiring “indications of interest” in purchasing these items “when they become available.” My theory of abundance is that when anything becomes ubiquitous - whether content or anything else – it drives the human desire for scarcity. As shoes are commoditized, we buy higher end shoes, etc. But what if our access to something special and scarce becomes less about affordability or waiting around after “indicating interest” and more about merit? Greg Isenberg called this “skill-locked commerce,” the idea of gating purchases based on abilities or some status achieved through merit — and the belief that the earned privilege to buy something will increase desire and activate purchases. What if we needed to solve a riddle or prove something to gain access to a particular good or service? What if you needed to run 20+ miles per week to get access to a certain pair of shoes? What if you needed a certain level of social following to get access to a premium collection? What if you could only watch a particular film in person for a limited time, perhaps only after watching the prequel and answering a set of questions? I think this is a fascinating area of human behavior and commerce to explore in the age of abundance.

Conformative Software, and the end of lowest-common-denominator user experiences.

There is a new variant of mainstream software I see on the horizon that could change consumer and enterprise software as we know it. While we’ve covered “DIY software” (where you can use products like Replit or Cursor to generate applications for all sorts of purposes) in previous editions of Implications, conformative software is different. Conformative software is a new wave of apps that become more tailor-made the more you use them. They are designed and built with a new form of accommodative user interfaces that adjust to your skill level, your needs, and your preferences. Functionality and complication can vary by user. This new variant of products literally conforms to each user, using AI to leverage data and preferences about the user to instantly generate and modify the interface, features, and user experience.

The greatest implication of conformative applications is the chance to FINALLY defy the infamous Silicon Valley/Alley product lifecycle: A simple product launches, users flock to the simple product, the simple product adds features (aka complexity) for power users and monetization, and then users flock to a different simple product. This seemingly endless cycle has fueled a pattern of apps replacing apps replacing apps for decades. But what if conformative generative applications end this cycle with applications that become more of what you want and less of what you don't…the more you use them? These apps could remain forever simple for new adopters, while getting more and more sophisticated for longtime users.

A great example just happened with the update of the Apple Photos app on our iPhones. Many of us iOS customers were dismayed by the added dose of complexity in a once simple Photos app. While some customers were probably satisfied by the new capabilities introduced around photo editing, organization, and collaboration, many others were confused. Herein lies the problem of developing software for more and more customers: It becomes increasingly difficult to make any one customer completely satisfied. When you have more customers with ever-evolving and diverse needs, and seek to remain competitive in market by adding more features, most products become overly complex. It’s actually kind of bonkers that applications require the lowest-common-denominator of features and accessibility to be successful. I imagine a world where app interfaces are no longer shared experiences, but rather hyper-personalized. As one X friend remarked when I first shared this idea a few weeks ago, “we’re gonna have to break in software like we break in shoes.”

Ideas, Missives & Mentions

Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, Adobe folks…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind (including the premium of “cooked compute,” feedback on cognicos, “personalization effects” as the new network effect, and a few other thoughts as a builder/angel investor). Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

(1) A new wave of investigative journalists will break news via betting markets. We learned in the election that betting markets are, in most cases, more accurate than professional polling and news reports of voting trends. While there seems to be no disincentive for spreading false information these days on social media, people act differently when money is on the line. Making a wild claim? Put money on it. Don’t think vaccines work, make the claim alongside a bet and we’ll see how you fare! The general population may benefit from bottom-line-driven fact checking and the use of betting markets to suss out the truth of what is really happening. New mechanics for reconciling the veracity of news (mutually agreed upon objective measures, etc) will emerge when money is on the line. I think we’ll see a new cohort of investigative journalists and researchers become wealthy news sleuths by sourcing proprietary insights and breaking the news via betting markets as opposed to mainstream news outlets that are funded via advertising/clicks. Further thoughts here.

(2) DIY software will revolutionize apps for consumers AND the enterprise. There has been much discussion of AI code reviews, GitHub co-pilot, and no-code application builders for the enterprise, but what are the implications of agent-assisted software development for consumers? Quick apps for your home or family were too hard to build until now. I think we’ll see some pretty remarkable and super niche software applications emerge in 2025, by and for consumers. And in the enterprise, the cost calculation of building your own internal tools will start to merit AI-made homegrown solutions to workflows and enterprise functions (and increasingly agents will replace these functions, per the last forecast on the list!) as opposed to the usual “find a SaaS product to solve every need.” Explored this further here.

(3) Augmented reality (AR) will continue to evolve on a slow boil in 2025, but will transform our lives as we know it within three years. Recent discussions of the impact of AR remind me of discussions about the impact of AI in 2016 - everyone knew it would be transformative, but couldn’t declare exactly how or when. Behind the scenes, the key pillars for AR to change the world are getting stronger - the devices are getting lighter (META’s Orion device is 96 grams), the 3D creation software is becoming both more powerful and more accessible (e.g. Adobe’s Project Neo, etc), the resolution of immersive experiences is getting better, and the role AI will play powering a new contextual operating system for AR is becoming increasingly clear. In 2025, we’ll see these fundamental pillars strengthen and more prototype devices emerge.

(4) Top AI talent will start shifting from working in the companies pioneering the latest and greatest AI to taking leadership roles in the industries that will benefit most from AI. Remember when most internet pioneers worked at Netscape, Prodigy, AOL, etc…until every company was, more or less, reimagined in the age of the internet? As LLMs and other generative AI models become increasingly open source and generally commoditized, the accrual of value will shift from those who make the AI to those who can make the most use of it. Certain industries where imagination and risk-taking can be most unleashed by AI will start to be reimagined by the top talent that can best apply these learnings and capabilities. Top of the list? Storytelling and filmmaking, privately owned hospitals, any customer-service intensive business, and every private-equity owned company (for which the “refactoring and re-imagination” playbook can be deployed top-down for massive ROI).

(5) We’ll see a new wave of small businesses emerge that use an AI-native tech stack to refactor their operations. Some of them will be small teams and large businesses, and others will be more artisanal yet (finally) economically viable. Indeed, 2025 will usher in an era of scaling a business’s reach and aspirations without growing headcount and expenses proportionately. But we may be surprised by the highly niche and crafted nature of these new businesses. Why? As every big company floods the zone of our attention with increasingly enticing marketing and cheaper and more personalized versions of everything, we — as consumers — will crave more scarce and authentically human experiences, often provided by a small business run by passionate people. In response, more artisanal-like and privately-owned businesses that may have been uneconomical to run before will start to emerge, powered and made economical by AI-driven tech stacks. Their products and their vision will be intensely personal, but the mechanics of their billing, marketing, etc. will be machine-driven. This trend should spawn massive growth in small businesses and address growing consumer demand for smaller scale and human-crafted version of everything. Discussed this more here.

(6) The rise of new "cognico" companies, designed to optimize compute-driven cognition as opposed to human productivity and reasoning. Cognicos will have data and compute at the center, with humans playing the unique role of stewards and orchestration designers and engineers as nodes. Until now, humans have been the reasoning layer of every organization. The next-gen company will be run by a combination of inference engines (real-time computational reasoning running every function and driving actions across the business), leveraging a variety of pre-trained AI models and a large amount of deep and proprietary data at the center of the company. This core of data, models, and computational reasoning will all be surrounded by “nodes" – the modern version of every “function” of a company from HR to product to sales - which, together, compose the logic layer that performs every process and operation of an organization. Perhaps some of these cognicos will also be decentralized, governed by DAOs (of people), and operated primarily by the cognition-driven stack? Old companies were designed to help people work efficiently together. New companies will be designed for cognition - the way a brain works. Here's a quick sketch of the way i've been thinking about this future construct of a cognition-driven company:

No doubt, 2025 will be another year for the record books.

Let's commit to playing with new technology to find its applications (remember, novelty precedes utility). Let's double-down on what makes us human - our empathy, our own stories and taste, our ingenuity, and our desire to explore different paths and attempt entirely new approaches to old problems. Let's keep sharing our ideas liberally to enrich collective thought. And let's be grateful that we get to live through another platform shift and moment of acceleration, and mindful of the responsibilities to keep asking the tough questions and being creative about what can go right AND what can go wrong.

Thanks for following along this year, and for all the feedback and dot-connecting along the way. Happy new year. -scott

• This edition is a bit “out there” (fair warning). We’ll explore forecasts and implications around: (1) the rise of cognition-driven companies (aka “cognicos”) and what I’ve come to call “the cognition stack,” and why the value of AI may ultimately accrue to the first cognicos in every industry. (2) how support, social, and sales are all conflating into the same channel and (3) some surprises at the end, as always.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for quality, density, and provocation vs. frequency and trendiness. We don’t cover news; we explore the implications of what’s happening. My goal is to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • Why better management and measurement tools — and offloading of mundane and repetitive work to AI agents — will increase the talent density in companies.

  • How “proof of human” behind a creation will become a signal of meaning, and other insights around “artifacts of humanity” in the age of AI.

  • Over the coming years, we will stop going to the sources of information as the perfect summary, answer, or solution is generated and surfaced for us by trusted AI tools. Let’s explore the implications for the news industry, every website, and software as we know it in the era of abstraction.

The Cognition Stack, the rise of AI-native cognition-driven companies (aka “cognicos”), and why the value of AI may ultimately accrue to the first cognicos in every industry.

Let’s fast forward a few years and imagine what a new company born in the post-AGI world might look like. Bear with me, as this will be an adventure.

In such a world, every company will ultimately become a compute + data-centric company. Until now, humans have been the reasoning layer of every organization. Sure, every company uses technology, but mostly for the purpose of driving productivity and analysis to help humans make decisions and take action. In contrast, the next-generation company will be run by a combination of inference engines (real-time computational reasoning running every function and driving actions across the business), leveraging a variety of pre-trained AI models and a large amount of deep and proprietary data at the center of the company. This core of data, models, and computational reasoning will all be surrounded by “nodes" – the modern version of every “function” of a company from HR to product to sales - which, together, compose the logic layer that performs every process and operation of an organization.

Old companies were designed to help people work efficiently together. New companies will be designed for cognition - the way a brain works. ChatGPT defines “cognition” as “the processes and activities involved in acquiring, processing, storing, and using knowledge. It encompasses a wide range of mental functions—such as perception, attention, learning, memory, language comprehension, reasoning, decision-making, and problem-solving…the ensemble of all the mental actions and processes through which information is perceived, integrated, and acted upon.” I’ve come to call this new type of company, with a stack of technology that looks more like a mind than a traditional organization, a “cognico.”

Today, we will explore how a cognico will function and the important role that humans will play (don’t worry – we do still have an important role!).

• Nucleus: The nucleus of a cognico is compute and data: AI researchers far smarter than I am have stated that the race to get the best “pre-trained” models is slowing. Once all of the world’s information has been leveraged to train state-of-the-art LLMs (thanks internet!), pre-trained models reach the limits of what they can do and the game shifts to the inference/reasoning layer. This reasoning layer is the real-time computing that leverages your own data and provides the logic to solve problems and take actions. In such a world, you can imagine that the nucleus of every company will become (1) the data unique to the company, (2) the LLMs and media models (trained on both proprietary and general data) that are used for various types of activities, and (3) the inference computation that powers the reasoning layer.

  

• Nodes: The functions of a cognico are “nodes.” The current functions of a company like Financial Planning & Analysis, Marketing, Sales, HR, etc will evolve to become nodes of computing logic that leverage the nucleus (data + AI) alongside APIs from AI-first start-ups emerging in each vertical - from legal to sales and beyond, and the deft AI innovation happening across incumbents. And each node will also be staffed by humans (which we’ll discuss in a moment). Every function-specific node will be governed by a specific set of objectives, goals, and rules. It is likely that OKRs (the “objectives and key results” framework used by many modern companies these days to measure performance) will reign in this world, but the notable addition to the mix will be rules. Rules will be a mix of AI reasoning-engines that drive alignment with broader goals and adherence with laws and acceptable practices, as well as human-designed and -imposed rules that help maintain brand standards and cultural principles. What will the humans within each node of a company do? Humans will serve the role of either stewards, orchestrators, or leaders, and they will all be stakeholders of the many nodes across the company.

• Stewards: The “stewards” of every node will ensure that the underlying models, the quality and integrity of the data, and the logic of their node is effective, aligned with goals, adherent to rules, and is constantly optimized. These node stewards will also be trained to develop and ensure alignment with OKRs and develop, implement, and ensure adherence to “rules.” Their skillset is somewhere between a Product Manager and a Program Manager. In a cognico, where the majority of functions and actions are only accessed via agents through the logic layer and run autonomously, rules will be very important. Rules for obeying laws and regulations across different geographies, rules for adherence to company policies, rules for brand compliance, rules for pricing and business management, and the list goes on. I expect to see “rules engines,” delivered in the form of APIs, emerge and become an indispensable part of the cognition-driven company stack (if you know of a startup doing this, let me know!).

• Orchestrators: An “orchestration designer” is the evolution of a product leader and designer for the AI era who works above the nucleus. An “orchestration engineer” works within the nucleus. I call these people “orchestrators” because they primarily, if not exclusively, work with components that represent AI models, APIs, and very complex prompts. They are conductors of logic, leveraging or instrumenting the various components in a cognico’s nucleus. An orchestration designer understands the capabilities of pre-trained AI models for their node in particular, as well as their own company’s APIs and third-party software components that are ultimately stitched together to execute and optimize the functions performed in their node of responsibility. An orchestration designer also understands the language for programming and testing the logic layer of nodes, especially when it comes to using a new breed of visual design tools, let’s call them “orchestration tools,” that build automated workflows using these various models, components, and OKRs. Most importantly, orchestration designers have great taste. They will ctry creative things to unlock new edges in the market or attract attention through meaning and ingenuity. Orchestration Designers will also develop and evaluate the subtleties of sales communications sent to prospects, marketing copy and brand assets that are shared publicly, or an HR-related message sent to human employees. They are wildly imaginative and able to explore all sorts of permutations of workflows, models, and components using orchestration tools. They are the final mile of human selection and decision-making that will ultimately help differentiate the output of every company competing to win in every industry. An “orchestration engineer” is responsible for setting up, maintaining, and optimizing the technical components of the company’s nucleus that relate to their particular node — the data, the models, and the inference that powers their node and its contributions to the organization.

• Leaders: I don’t have a better name for the role that a group of people will play setting the goals and direction for every company. Leadership will be more important than ever before in the era of cognicos. If the company itself is a brain, the leaders must be the heart. They must set the overall goals, they must include leaders of each function charged with thinking about the future of each function in a creative way that models “trained on the average of what’s been done before” cannot. They must gut-check every decision, they must declare the double-bottom line for doing good in a world that tends to optimize for profit. They must also prioritize the parts of business that aren’t intended to scale - the art of business. The leaders of cognicos must also identify and advocate for the very human activities that will distinguish the company’s product and brand in the marketplace. While computers will help us think, they struggle to make us feel. Leaders will be responsible for the humans and the humanity of modern organizations in the age of AI.

• Every company will find, engage, and serve its customers via agent-based experiences and contextual UI powered by AI. This “logic layer,” powered by agents and user interfaces tailor made for specific needs, is the ultimate interface layer (an obsession of mine since 2014). If your customers are consumers, it is likely that your service will be discovered and used through OS-level agents, AI-enhanced browsers, or specialized interfaces that consumers use to plan and manage their everyday lives. If your customers are employees in companies, your services will likely be delivered as agents within customer workflows, or perhaps new interfaces to conduct work that don’t even exist yet. Future customers, whether they are consumers or employees in a company, will summon and use AI capabilities at the agent-driven surface of their daily lives and work. Will the mainstream operating systems of today reign in this new world? Will new agent interfaces emerge that we will all use in certain contexts? We shall see.

• What are the implications for the first cognico companies in every industry? Quite simply, the value of every industry will accrue to the companies that have the highest speed and quality of transformation into a cognico. The most important question as an investor these days, or a founder or executive making a career decision is: Going industry by industry, how much refactoring is there to be done? Such refactoring could occur in customer service costs, logistics costs, production budgets, sales overhead, and the list goes on. While today we see top technical talent wanting to work at the major LLM model companies, I think we’ll see the next wave of top talent flock to each industry to help drive these transformations. No doubt, the major LLM model companies are rapidly improving AND undercutting each other in price, accelerating the path to commoditization — especially given the rise of local and open- source models. What if too much attention (and capital) is being deployed to companies that MAKE the AI models while the majority of value accrues to the first to who use them in each industry? As AI completely refactors and reimagines the way companies — and entire industries — operate, I believe those who use this new technology may experience the greatest benefits.

• What are the implications for business models? We’ve explored in a previous version of Implications the rise of outcome-based business models, as well as paying for agents like you pay for headcount. No doubt, business models will transition in the rise of cognicos. Instead of paying for seats, we will pay for generative credits (essentially marked-up compute, or shall we call it “cooked compute?”), outcomes, and performance. Instead of charging for time spent, we will charge for expertise by project and the specialized compute required to meet our goals.

• What are the implications for humans? Consumers will get better prices, make better decisions, and ultimately benefit from better experiences in the age of AI. But what about the workforce? How many leaders, stewards, orchestration designers and orchestration engineers will companies need? What about the next generation of talent? As companies become cognicos and refactor their workforce, growth-oriented companies will have a choice of either shrinking the employee count or taking on more: more projects, more products, more campaigns, more testing, more high-touch human customer service, etc. My guess is that people will be retrained and redeployed to participate in more nodes to deliver on the growing aspirations of growth companies. As for the companies that choose to stay small and efficient, we will see 10x or 100x more of these companies. There are so many products and services that people would want but are far too niche to be built and deployed in a profitable way. We’re starting to see very small services emerge, from research and consulting to career support and production shops, that are founded and led by just a few people, without the intention of ever scaling. Finally, let’s talk about the return of crafts. In the era ahead, humans will crave more scarce, authentic, and offline experiences than ever before. We will crave small restaurant experiences with proud chefs. We will crave one-of-a-kind art infused with human story. We will crave theater and emotional films with deep meaning. We will crave shared experiences and live music. In the age of AI, there will be rampant demand for stuff that only humans can create.

Social is support is sales is social.

What’s becoming increasingly clear to me is that three formerly different stacks of products and people in companies of all sizes are increasingly becoming one. When it comes to social marketing, every event is a public marketing moment - whether it’s a customer complaining, a product leader responding to a question, or a social post marketing the product. As a result, the support folks need to start thinking like marketers, and vice versa. Same goes for sales, especially in the age of AI Agents that will greet customers across every digital experience with a “‘how can I help you?” If a customer asks a question, does that go to support or sales? It shouldn’t matter — every support and marketing moment is also a sales moment. In the social world, whenever a customer complains, everyone else is watching the response and judging the brand accordingly. Implications? As these three functions collapse into one another, a new breed of products — and possibly organizational designs — will be required to manage and deliver world-class experiences. New measures must also emerge that quantify these channels crossing one another, and help optimize them.

Ideas, Missives & Mentions

Hope you have a restful holiday and new years.

For those ready for an extra dose, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, Adobe folks…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind (including two ideas I hope someone makes happen, the consequence of NOT taking creative risks, reconciling memories and dreams, and a periodic table of branding). Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more

• This edition explores forecasts and implications around: (1) Where betting markets may take news, (2) DIY software revolutionizing consumer and enterprise applications, (3) true art standing out, (4) the decentralization and community programming of money, media, and more, and (5) some surprises at the end, as always.

• We have Part III of Insights for Modern Product Leaders.

• If you’re new, here’s the rundown on what to expect. This ~monthly analysis is written for founders + investors I work with, colleagues, and a select group of subscribers. I aim for density and provocation vs. frequency and trendiness. We explore the implications of what’s happening to ignite discussion, socialize edges that may someday become the center, and help all of us connect dots.

• If you missed the big annual analysis or more recent editions of Implications, check out recent analysis and archives here. A few recommendations based on reader engagement:

  • When is augmented reality going mainstream? A review of progress.

  • Transcending prompt-based creativity with the “controls era” ahead.

  • The more edgy implications of the transformation of entertainment.

  • Making sense of rapid displacement and “the best model” rat race.

Assertions & Implications

Lets jump into five assertions with implications worth noting — all around the theme of unexpected developments of new tech and things top of mind for me.

• DIY Software will revolutionize apps for consumers AND the enterprise.

• The next wave of investigative journalists will break news via betting markets. Is truth best surfaced with incentives?

• True art lies outside the distribution curve of what’s been done.

• Insights for The Modern Product Leader Playbook, Part 3

• Bonus section: thoughts on Bitcoin as an elaborate and decentralized form of Universal Basic Income (UBI), and more.

DIY Software will revolutionize apps for consumers AND the enterprise.

Replit (am an investor) has been blowing my mind lately. With a simple prompt, I was able to make an app (simple but functional!) for my kids to track their book list and daily reading goals. Describing my idea with a sentence kicked off a form of “AI product management,” and then, with a click of a button, Replit wrote the code, developed the logic, created a database, and deployed server space. Within two minutes, I had a v1 to start playing with. Wow. There has been much discussion of AI code reviews, GitHub co-pilot, and no-code application builders for the enterprise, but what are the implications of agent-assisted software development for consumers? What if there are billions of people making apps for themselves? What “picks and shovels” tools, APIs, security services, and other hosting and deployment tools will prove essential in such a world? I have the same questions for the enterprise: Will the cost calculation of building your own internal tools start to merit homegrown solutions to workflows and enterprise functions as opposed to the usual “find a SaaS product to solve every need?” Will new “orchestration layer” homegrown technology stitching together various AI models into intricate workflows replace whole swaths of Enterprise SaaS products?

The next wave of investigative journalists will break news via betting markets. Is truth best surfaced with incentives?

Many of us spent the last year focused on betting markets like Polymarket and Kalshi, with a keen eye on the gaps between these anonymous financially-driven projections and professional political polling. Of course, the betting markets proved far more accurate, which begs the question of why and whether the implications will spread far beyond politics. One illustrative story is of an independent researcher in Europe who leveraged “neighbor polling” (small polls where people are asked who their neighbor is voting for, as opposed to their own voting intentions) to discover a gap between those results and the official polling of registered voters. This researcher gained enough conviction to reportedly bet $30M on Trump winning the election. In essence, this was an independent researcher who leveraged proprietary insights to influence the news “Trump is going to win” through market dynamics as opposed to news outlets. With the rise of decentralized betting markets, will the people who break the news (or perhaps even make the news?) start profiting from the news? Will we see a new cohort of investigative journalists and researchers become wealthy news sleuths by sourcing proprietary insights and breaking the news via betting markets as opposed to mainstream news outlets that are funded via advertising? Does any of this constitute “insider info” if those placing bets to break news are, in some way, a part of the story they are breaking? Needless to say, we are in the early days of betting markets. However, as we consider ways to restore “truth” and trust in the news, I wonder if betting markets can be part of the solution? Perhaps there will be a news site with a series of claims (headlines and articles) where people can “bet” on the validity of these claims. And then an independent bipartisan group of people can surface and source facts that, collectively, serve as the arbiter of who wins and loses the bets? While there seems to be no disincentive for spreading false information these days on social media, people act differently when money is on the line. The general population may benefit from bottom-line-driven fact checking and the use of betting markets to suss out the truth of what is really happening.

True art lies outside the distribution curve of what’s been done.

What is art? As a builder of products for creative people, someone who spends a lot of time working with leaders in the AI space, and as a board member at the Museum of Modern Art, I do think a lot about art in the age of AI. There is a general misunderstanding that “AI can make art” when, in reality, AI is a tool - much like a camera or brush - that someone can use to make art. And, like most photographs or brush strokes, AI is most often creating something “average” and often common. The potential of a tool to make art is up to its user. Much like any kid with crayons can draw a tree or anything else they’ve been exposed to, GenerativeAI can conjure up a tree if you ask for one, but it will come from the commonality of attributes across the trees in the model’s training data. Under the hood, large language models (LLMs) and generative media models are trained on massive amounts of data to help predict a sequence of words or pixels based on the prompt you provide. So, if you prompt for a tree, you’ll essentially get the average representation of a tree. You won’t get the most exotic tree that resembles one-of-a-kind trees that lie outside the distribution of trees…unless you sufficiently describe such a rare variation of tree. Which takes ingenuity. The point I am trying to make is this: If AI is trained on the masses of data, then the responses and capabilities of these models “emerge from the center of the distribution of inputs,” as one friend framed it. These models are designed to give you “what you would expect.” This is the opposite of art, which engages you by being what you didn’t expect. Which brings us back to the question, what is art? Art exists outside the distribution of what’s been done. Once you copy art, it is less art. But when you’re inspired by art and inject your own humanity and unique perspective, you’re liable to make art. I am reminded of producer Rick Rubin’s view on artists: “Artists allow us to see what we are unable to see, but somehow already know…the artist’s perception reminds us of who we are and who we can be. The personal is what makes art matter. Our point of view, not our drawing skills or musical virtuosity or ability to tell a story.” With this context, as the tools change and skills become more accessible to more people, the unique point of view we each bring to our work — much more than the tools we use — is what makes art. Art comes from the edges, outside the distribution curve of what’s been done (aka the data).

Insights for The Modern Product Leader Playbook, Part 3

Continuing this monthly collection of insights I share with or learn from modern product leaders I admire. I’m pulling out 3-5 of these from my collection for each edition of Implications, given many of you are founders or leaders of various functions or teams across the creative and technology industry. Pretty sure I’m done writing books, so figured this was an optimal place and audience to share these insights…

• Collapse the talent stack. A lot of the magic I’ve observed in teams over the years happens when the talent stack is collapsed - when a designer also codes, when an engineer has a growth hack skill set, when a product leader is great at copy. Vertically integrating interdisciplinary collaboration in a single human is f’n awesome, and you should transcend traditional org structures of teams and make exceptions for who you hire when you encounter an opportunity to collapse the talent stack.

• For early-stage products, density of usage matters more than number of users. For consumer products - or any product with a network effect - one lesson and observation is just how critical small pockets of engaged users are for organic growth. Small pockets of highly engaged users help a small number of people discover the full value and utility of a product far faster than depending on mass adoption. Facebook went college campus by college campus with lots of intentionality, so as to ensure small pockets of high engagement. Pinterest also did this with small networks of middle-American moms. And I remember Periscope thriving with > 50% DAU in a small group of people within Silicon Valley. Most recently, a new AI-driven news app called Particle launched, after an extensive public beta designed to demonstrate high density of real usage. In fact, I’d argue that any product that relies on achieving scaled use in order for users to feel successful is doomed to fail. Startups, given their reliance on small pockets of deeply engaged users, have an inherent advantage here. Big companies tend to get lost in large numbers and fail to optimize for dense nodes of users. The best path is high density of usage in small pockets, and any company in growth mode should (ironically) focus on density of usage pocket-to-pocket over total number of users growing person-to-person. A pocket of dense usage demonstrates both engagement and the network effects required to grow.

  

• When you love what you do, you get a lot more done. Initiative outperforms experience. Your job as a leader is to align people with tons of initiative most closely with their area of genuine interest. When it comes to hiring - especially for early stage ventures and bold turnarounds - initiative outperforms experience. This means you don’t want to be a resume snob, you want to hire people who have demonstrated, over the course of their lives, the willingness to take initiative in anything that matters most to them. If they loved sailing in college, they started a sailing club. If they are an avid reader, they co-chair a book club. These are all signs. Also, I’ve always firmly believed that a labor of love always pays off, just not necessarily the way you’d expect. Hire accordingly.

Assortment Of Findings & Shout-Outs

• Dystopia or Utopia? No surprise, I like when technologists take the pen and think out loud about some of the tensions around new technology. On this topic, Vinod Khosla recently sent me his essay aptly titled “AI: Dystopia or Utopia?” He managed to transcend the general myopic soundbites by breaking down  the major areas of concern and opportunity. Most naysayers believe the negative implications are inevitable. But he points out that they are a choice. I would also add that humans tend to evolve to account for and protect against the negative implications of new technology rather than fall victim to them. We created traffic laws and issued drivers licenses to protect against the danger of cars, etc… The accommodation of change is part of the human condition. Vinod also addresses some of the concerns I also share around rising income inequality and AI leaving groups of humans behind. The Norway "transition funds" analogy is one I often cite and I wonder how we can get politicians to grasp this. Capitalism certainly needs to evolve. As a father, I think about this often as a preventative measure against civil unrest from inequality (probably one of the actual risks we should be most focused on!). Check out his essay.

• Making it easier to bring your product to the enterprise. Just a bit of a shout-out to WorkOS, a company I have watched evolve a bit over the years that helps startups start selling to enterprises with just a few lines of code. So many products I have supported over the years have struggled to scale by engaging enterprise customers, and WorkOS is finally change the game with a complete user management solution along with SSO, Directory Sync (SCIM), and Fine-Grained Authorization (something many enterprise customers require!). Much like products like Stripe, etc, it is self-serve and easy to integrate in minutes with modular and easy-to-use APIs. They now power hundreds of high-growth companies including Perplexity, Vercel, and Cursor.

• What other centralized parts of life will become decentralized and programmed by the people? With the rise of Bitcoin, there is no shortage of hot takes. But one consistent voice in this arena is Balaji, who noted recently “All media became social media. All money becomes cryptocurrency. I was struck by the theme of things once controlled by small groups of people in back rooms becoming programmed by the masses, whether it be those programming a news channel being replaced by social media and community curation, or the small group debating federal reserve policy becoming replaced by a decentralized and programmable form of money. This prompted the question of what other parts of life are poorly centralized today? Perhaps healthcare, where we struggle to access our own data and await very lengthy FDA processes to get ahold of new drugs and treatments? Perhaps education, where we conform to a curriculum as opposed to the curriculum becoming tailor made (perhaps using AI tutors) for each student? We also can’t always assume that “programming by the people” or the wisdom of crowds is always better. One might also argue that mainstream media becoming social media programmed by the people could exacerbate the less wonderful natural human tendencies of people, like being captivated by conspiracy theories and being manipulated by algorithms designed for engagement more than truth seeking.  No doubt, this theme of “collective programming” of things that were once centralized will be a dominant theme of the decades ahead.

Ideas, Missives & Mentions

**Finally, here’s a set of ideas and worthwhile mentions (and stuff I want to keep out of web-scraper reach) intended for those I work with (free for founders in my portfolio, Adobe folks…ping me!) and a smaller group of subscribers. We’ll cover a few things that caught my eye and have stayed on my mind (including an unorthodox take on Bitcoin and universal basic income, data provocations related to toys, and more). Subscriptions go toward organizations I support including the Museum of Modern Art. Thanks again for following along, and to those who have reached out with ideas and feedback.

Read more