The Rise of Open-Sourced R&D, How Communal Resourcefulness Will Protect Us, & Wild Data Provocations
In this edition, we explore how the collective will use breakthrough technology to outperform R&D departments, security companies, and bad actors. How do we leverage the masses in the age of AI?
Edition #15 of Implications.
2024 will be flush with change and fascinating implications. We face a pivotal year ahead amidst platform and societal shifts, breakneck speed of innovation, dust settling from economic volatility, and material changes in human preferences and desires (within and beyond the workplace). This ~monthly exercise exploring the implications continues. It’s been fun to share and get your input. Just one year in, this analysis now reaches over 26k readers — including creative minds, investors, builders, and leaders of companies I admire — spread through word-of-mouth. Thanks for the accountability and dot-connecting.
Jumping In: If you didn’t catch it, I shared a set of forecasts and implications top of mind for me a couple weeks ago, check it out here. This edition of Implications explores forecasts and implications around: (1) the rise of open-sourced R&D, (2) the concept of communal resourcefulness and how AI trained on our collective lessons-learned will be our greatest defense against AI misused, and (3) some data provocations + other surprises at the end.
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 recent editions of Implications, check out recent analysis and archives here. A few highlights include how algorithms are polarizing us and what we do about it, joyspan vs. healthspan vs. lifespan - and other wild concepts bound to become commonplace, and the most-shared edition of the year around the coming wave of personalization. OK, let’s dive in…
The Rise of Open-Sourced R&D
The function of traditional R&D in every company is to come up with new ideas — whether new features or entirely new products. This work is typically performed by a specific group of people charged with idea exploration and some form of validation in the form of customer research. But is the assumption that the best ideas always come from within a company outdated? Innovation within companies is inherently compromised by antiquated assumptions, bureaucracy, and the gravitational force of quarterly goals. Why wouldn’t a passionate and obsessive customer have insights that salaried employees fail to reach on their own? What if better ideas for new features or partnerships are abundant, but the friction to explore them and influence companies was always too great…until now?
The next generation of product development and brand partnership exploration, especially among consumer brands, will increasingly be done by customers using AI creative tools, validated on social platforms, and then executed by companies. It starts with a clear vision for something great and the impetus to act on it. Until only recently, the “act on it” part was incredibly expensive and time-consuming. But now, as the tools for idea expression are accessible and commoditized thanks to generative AI models and easy-to-use web-based tools, there is very little friction for expressing an idea visually. Once you take a few minutes to express your idea visually and share it, social media provides a platform to see whether others engage and amplify. Once something catches on, brands can leverage this social traction as a form of validation and bring the idea into their advanced R&D pipeline. For instance, let’s say you have an idea for a collaboration between Lego and luxury handbag maker Louis Vuitton. This recently happened and spawned an internet sensation with countless people asking where and when they could purchase this fantasy collaboration. Word on the street is that both Lego and Louis Vuitton have taken note…
We’re entering an era in which our customers will not only show us what they want, but actually drive the first mile of R&D for us. New lines of clothing, new features for products, new genres using iconic storylines - these ideas have always been everywhere but only had a chance of being actualized if they came from within. But the world seems to be changing due to the following forces:
Increasingly powerful imaging models with the ability to customize and constrain. While the legality of playing with the brand collateral of a company you admire is questionable, the ability to engage customers and fans with the R&D process using actual brand assets and product parameters is now widespread. The niche world of “fan art” for brands has the opportunity to become a sanctioned R&D initiative. Companies could invite customers to leverage “fine-tuned” GenAI image and video models that were trained with the brand’s copyrighted material. For instance, if people love making memes (or their own stories) with Star Wars characters, why not view it as a new genre and enable sanctioned GenAI Star Wars models with some guardrails (and new monetization models)? Over the coming years, we will see more hypothetical products and brand collaborations than real ones. The question is, will companies fight it or leverage it?
Community curation with the “credible mass” overpowering the “critical mass.” Of course, the growing surplus of ideas and concepts brings the challenge of finding the gems. How do we surface quality more effectively? How do we compensate and incentivize open-source R&D? While social platforms are proving effective at pressure testing and gauging the customer response for new ideas, sometimes the general public isn’t the best curator for what’s next. It’s not just about the critical mass (how many people like something), it’s also about the credible mass (who likes something). As algorithms mature, I suspect social platforms will make a new set of sorting and filtering tools available for enterprise customers who want to sort through content in new ways. These algorithms will weigh the influence of people differently based on their expertise and area of influence. For example, as you sort through ideas, you can allocate more influence to curation by established artists. Brands want to know what “tastemakers” prefer, and social platforms have an opportunity to help brands differentiate between the credible mass and the critical mass.
Taste > Skills. Open-sourced idea generation and tapping the ingenuity of anyone beyond your marketing or product development departments prompts another question: Perhaps taste and ingenuity are far more important than skills in the age of AI? Taste seems more scarce these days, and is an increasingly differentiating trait as skills-based productivity is offloaded to compute. This realization makes me contemplate new structures and incentives that companies should use as they hire the next generation of talent or devise ways to engage tastemakers. I also wonder; What are the basic elements of taste? Is it the amalgamation of life experiences and what we’ve been exposed to? Is taste related to our state of mind and perspective on the world? How do we develop taste in the next generation? Finally, we must realize: Having taste and leveraging AI is our ultimate defense against being replaced by AI.
Communal resourcefulness, in the form of shared block lists and threat filter models, will foster a more secure future.
The greatest risk in the AI era is not leveraging AI to mitigate threats from bad actors using AI. When people ask me what scares me most about AI, my answer is the technology falling into the wrong hands without good actors having superior AI for threat mitigation. For instance, let’s consider scam prevention. In an era where phishing calls will feature your mother or friend’s voice, we will need some powerful defensive technologies to stay safe. A big part of the solution will be AI in the form of agents deployed to protect our interests. I believe these agents will leverage models trained to detect bad actors. But another opportunity I am really excited about is the prospect of all of us banding together and keeping an eye on each other in the form of shared “block lists” for scammer phone numbers, email address, and social accounts. For this to happen, AI will need to be trained on data from a broad network of participants.
Voltron, via AI. The risks we face as individuals are best mitigated by working together. While this is an old underpinning of society, we have yet to capture all the lessons learned (often the hard way) that occur at the edge (on our devices, in our daily lives). Whether it is a scam call, a malicious email or text, a social media post from a bot, or some other deceptive signal, we typically delete or endure silent shame when we click. These one-offs, in aggregate, are an incredible dataset and shared resource that we have yet to fully leverage. Companies like Block Party and others are exploring this space. I am most excited about platforms aligning around open standards and portable datasets that can help train AI to outmaneuver the bad actors. This form of “community resourcefulness” is essential to learn from the vulnerabilities faster than the vulnerabilities become best practices for bad actors. For this to happen, platforms and people must agree to collaborate. I’ve learned this first-hand with the Content Authenticity Initiative — an open source effort my team helped start that now has over 2000 companies and organizations involved. Our objective is to have cameras, tools, and AI model add “content credentials” to assets so we know who made them and how. Our hope is that media companies and social platforms will start to show the credentials for media to help us verify where content came from and discern what we can trust. We need more models like this where the collective bands together in the pursuit of safety.
Community-sourced notes for every digital experience: One idea I’ve had over the years was building a layer across every website in the world that would allow visitors to leave a form of “community note” with suggestions, notes, and reviews that would be surfaced to other visitors (or perhaps just people you know). This layer above every digital experience could also include attribution information about the developers, designers, and AI models that made these experiences. When I invested in The Browser Company (maker of Arc) I wondered whether such an attribution and community sourcing layer could be applied to every website at the browser level. But I know Josh and team are pretty busy with the rest of their agenda. Either way, I hope some company makes this happen and that we leverage this collectively to better protect and inform each other.
Immersive experiences in the era of spatial computing offer a new dimension for community-driven guidance. As we build out this system of collective comments and warnings on the internet, consider how much more effective they’ll be as mixed reality and devices like Apple’s Vision Pro go mainstream. Sure, “reviews” on websites like Yelp are nothing new, but when these notes and pointers literally obstruct our view they’ll have far more power. I love the idea that we will curate the world for each other - whether for our friends or for everyone, and these collections of crowd-sourced guidance gain a new level of prominence as immersive experiences go mainstream.
Decentralized datasets become the most trusted community-constructed assets. The problem with the vision I’ve articulated above is the inherent bias and power of any one entity that controls this information. What if some spammers pay to get through on some “allow list” despite negative user reports? What if one company decides to monetize this power too aggressively or limit it to one platform? Is any guidance truly objective and forever reliable if it is controlled by one entity with its own incentives? For this reason, I am bullish on the prospect of the blockchain helping power such community-built mechanisms.
Ideas, Missives & Mentions
Finally, here’s a set of ideas/thoughts (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 stayed on my mind (notably a series of data provocations that speak volumes), as well as things I’ve learned as an investor. Subscriptions go toward organizations I support including the Cooper-Hewitt National Design Museum. Thanks again for following along, and to those who have reached out with ideas and feedback. Now, here are some worthwhile things to note:
Join Implications, and subscribe to continue: