The Personalization Wave, A Surge of Wildly Human-Intensive Non-Scalable Experiences, & Ideas Of The Month
We cannot surf waves without knowing where they may take us, right?! Let's discuss implications of the personalization wave and the anticipated rise of non-scalable experiences.
Welcome to Edition #6 of Implications.
This edition explores forecasts and implications around: (1) the massive wave of personalization coming our way, and (2) the coming surge of wildly human intensive and non-scalable experiences that become economically viable in the age of AI. And (3) (for folks I work with and subscribers) we’ll close with a few ideas, missives, and mentions - including one particularly wild idea called “Peanut Gallery.” The pics in part II are from my two weeks in Japan, and accompanying inspiration for my outlook on the rise of non-scalable experiences.
If you’re new, here’s the download on what to expect. This ~monthly analysis is written for founders + investors I work with, and a small group of subscribers.
If you missed the big annual analysis (shared more broadly), or more recent private editions of Implications, check out recent analysis and archives here.
Also FYI - had the opportunity a few weeks ago to discuss AI as well as some reflections on all that our team has been up to at Adobe over the last few years with Stratechery with Ben Thompson, some of which was initially explored in this newsletter. Ok, let’s dive in.
The Personalization Wave: The future of search, commerce, and decision-making will be hyper personalized through interfaces that cater to your style and preferences.
The case for personalization is my general thesis of breakthrough consumer technology: it takes us back to the way things once were - the ancestral comforts we long for - but with more scale and efficiency. In the ancient world of small tribes and towns, we were known. We long for the days when our local storekeepers greeted us by name and our neighborhood restaurants remembered our likes and dislikes. We want to be welcomed, served, and remembered for who we are - and valued for our loyalty. For example, I’d be fine with every restaurant in the world knowing that I am a vegetarian. I’d be fine with every e-commerce store knowing that I wear a size 9.5 mens shoe. However, I want to know HOW they know this information. I don’t want my data to be sniffed or scraped, I want control.
Over the past few centuries, as commerce and hospitality scaled through cities and technology, personalization was left behind. We became "logged out" visitors in all parts of our lives, greeted with generic menus and generalized calls-to-action. Today, we experience the lowest common denominator of digital experiences on a daily basis, and are oblivious victims of data collection and retargeting campaigns. All this, despite the fact that brands thrive or die based on conversion and retention.
Well, here’s a bold vision: the future should be personalized to your preferences. Every digital experience should be personalized for you based on the information you want brands to know about you - always in your control. E-commerce websites should welcome you by name, use your preferred language, know your gender, preferred sizes, colors, and ages of your kids. Restaurants should know your allergies, favorite dishes, and dietary restrictions. Hotels should know your preferred room temperature, and whether you have rewards status at other hotels or airlines, or if you are an influencer. Future AR and mixed reality experiences should be hyper-personalized, including your favorite colors and personal tastes and interests. Your preferences should be managed by you and shared purposefully, not deduced or purchased from others. How might recent advances in AI make this world possible, and what are the implications?
Step one, sync yourself. On the consumer side, you’ll have the opportunity to sync yourself (a personalized model trained off your data aggregated from a variety of sources) with any application or experience you engage with. Now, you won’t want ALL of your data to be used by EVERY application, so the killer technology here will be an intelligent arbiter - an “agent” of sorts - that acts on your behalf to personalize every experience as appropriate (and learns from what it gets wrong). Your model, and the enterprise models behind every experience in your life, will have a real-time agreement or negotiation of sorts to personalize your experience while optimizing for your interests and privacy.
Personalized “agents” will cater to your needs over the course of your life and become the dominant interface for your decisions. I liked how Patrick OShaughnessy outlined some of the possibilities from having a stable of personalized AI agents. In his view, we’ll have many specialty agents like a security agent that “helps you manage digital and personal and physical security,” and a business intelligence agent that “helps present most critical business intelligence in your organization.” I agree there will be many specialized models for the different functions of our work and lives, however I think our experience of these models will be aggregated to a few (if not one) major interface - if only for the benefits of having these models (and their datasets) work together. I can imagine each of us choosing an avatar and personality for this sentient being that ultimately becomes a life long relationship for each of us, helping us navigate the most inconsequential (which shoes should I buy?) to the most critical (who should I marry?) decisions of our lives.
“Recommendations” kills “Favorites”: No doubt, AI has reached the point in some verticals where it knows our taste better than we do. In such a world, the model’s recommendations may please us more than our own favorites. This is a forecast I have been brewing since I first shared in December 2021 when I stopped compulsively saving playlists that i discovered and loved on Spotify and fully surrendered myself to the algorithms. It was really the new “enhance” feature that did it. I now know that any song I like leads to a playlist i’ll like, and any playlist I have will always dynamically evolve and get better. Spotify’s new “DJ” feature also does this. In my world of music, recommendations have started to take the place of favorites. Where else will this happen in our daily lives? Will your favorite travel experiences suggest recommendations that transcend a Google search or a travel agent? Will the past fonts you’ve used as a designer, coupled with whatever is on your canvas, suggest the font you should use for a particular project. I think we’ll look back and realize that “favorites” were always quite limiting. “Favorites” has long been the way we, as humans, ensure that every circumstantial discovery and experience we love is able to be experienced and encountered again. Whether it is a favorite song, a favorite restaurant, or a favorite hotel…all these places and creations we discovered by accident ended up consuming our attention at the expense of new and (by the law of probability) even better experiences we’re decreasingly likely to discover during our limited time on earth. So, I welcome the era of AI to tap the collective actions of everyone with similar interests to make sure I don’t miss something I’d love. And yes, I will continue to add in a dose of randomness for good measure. Recommendations are on a path to not only become 10x more powerful, but also far more ubiquitous. At what point does your shopping cart automatically populate and you just remove things instead of add things? At what point do e-commerce websites give you an ever-optimized TikTok-like experience rather than categories?
The benefits of personalization compound over time. Many of us have already experienced the benefits of ChatGPT remembering our previous questions. I am now thinking a lot about the compounding benefits of a relationship with a personalized ML model after months if not years of working together. Does the model discover our quirks and biases? Does it develop an attitude or degree of sarcasm in order to engage us on our own terms? And what does this mean for marketing and the world’s largest brands giving their customers a more personalized experience? At Adobe, we’re thinking about how to help some of the world’s largest brands deliver personalized experiences to their customers. Today, most personalization is done by "segmenting" customers, but perhaps every brand will have a fine-tuned model for their brand as phase 1, and then a fine tuned version of that model for every single customer as phase 2? Prediction: average retention for most brands will be dramatically stronger in just a few years.
Non-portable data will be the ultimate moat. As many of you dear readers are founders, builders, or investors I work with, I am sure you’re also considering: how does this all play out and which products will win? I am focused on three vectors of differentiation with my investor mindset and as I lead strategy, design, and emerging products at Adobe: (1) is there a unique dataset - or proprietary understanding OF a dataset - that yields an advantage?, (2) is there a significant distribution advantage or a final mile advantage (like critical adjacent tools in a given workflow)?, or (3) is there a revolutionary interface advantage that, for some reason, would be difficult for others to deliver? But on the first one, I had a back-and-forth on this with one of the founders I work with, Dev Nag, CEO of CtrlStack, who made the point, "Proprietary data that is not portable can be a leverage point for training models with superior performance -- examples would be user click data, user prompt history, even instrumented system data that's not visible to (or charged to) customers. I predict that we've hit peak openness on the web, and we're going to see a retrenchment into more private data over the next few years as people realize how much of their public work can be easily co-opted by transformer-based models." Perhaps we’re also entering a new era of instrumentation to build these very native and non-portable datasets to truly transform user experiences? Dev went on to make the case further, “the hyper-personalization angle is absolutely inevitable -- the cost of inference and fine-tuning of these models on that private/personal data is dropping dramatically each month. Even a few months ago, we thought we could only do GPT-competitive inference on giant GPU server farms, and then llama.cpp came out and showed that you could do inference with a 65 billion parameter LLM (sourced from Facebook) on a Mac laptop, CPU only. 🤯 And then naturally, a bunch of folks built on top of those approaches, like alpaca.cpp. And people are already pointing out some relatively easy wins to make the performance and resource usage even better, like new bit quantization approaches. There won't be a technical barrier (for very long) to everyone having their own personal LLM trained on all of their data and activity, and likely multiple LLMs for you based on different domains of your life, different privacy realms (eg, this model is shared with my extended family to help plan group vacations and give gift ideas, this other one is just for me), etc.” Suffice to say, the differentiator in the future will be less about who had the largest public dataset or who spent the most on compute, it will be who built and made innovative use of proprietary sources of data.
A surge of wildly human intensive and non-scalable experiences are ahead of us as they become economically viable in the age of AI.
AI refactors previous use of humans, and unleashes new use for humans. Two weeks traveling Japan proved to be the perfect setting to contemplate some of the sweeping changes facing our society over the coming years and decades. The smart people I know generally agree that 80% of the work of 80%+ of jobs will be refactored significantly by AI. And I’m starting to see this as the current version of GitHub co-pilot increases engineering productivity by more than 30% (according to some early embracers) and all customer service oriented functions are already being transformed. And it’s still early days! So, the question that keeps me up at night is, what are us humans gonna do with all of our newfound time? Which brings me back to Japan, and this quaint Kyoto restaurant I found myself sitting in one evening. There were 10 seats, one chef/owner and one apprentice, and the most incredibly crafted experience. It wasn’t expensive, but everything was intentional. I found myself admiring this sensational and remarkably unscalable experience. The chef seems to make a good living, loves meeting interesting people, and gets to be wildly creative (the selection of glassware, the decor, the care and craft applied to every dish). Japan is full of these experiences, where art, curiosity, and craftsmanship yield tiny scattered wonders like “owl cafes,” micro arcades, plastic food shops, cotton candy shops, and the list goes one. I found myself wondering, why aren’t there 1000x more of these experiences in all societies? Why must the purpose of business be to scale, grow bigger, become franchises, squeeze in more seats, and compromise quality for automation and reach? Could a fundamental change in society, like mass automation and AI, spur both the growth and demand of human-intensive highly crafted unscalable experiences?
The rise of the experience economy will transform what we do with our time and for a living. As human workflows are refactored by a step-function and our capacity is liberated, I see a compelling future where the demand for and economic viability of crafted non-scalable experiences transforms society. The “experience economy,” is already underway with the emergence of experts for everything - from lactation consultants and pet therapists to for-hire violinists and chefs and…who knows what’s next. I enjoyed this post by Dror Poleg where he forecasts a world where “most people no longer need to work. Our survival depends on convincing them it's ok to do something else.” “There are many more professions to invent,” he declares, “and they will only be invented if more people experiment.” I believe humans instinctually crave scarcity and preciousness. We will want rare experiences, and will seek highly crafted versions of every basic and luxury need. With this demand, where is the new supply? I am seeing this in NYC as commercial real-estate is repurposed from stores to experiential installations. Whether it is Genius Gems - a space where your family can make things with Magnatiles by the hour, the trend of modern museum/instagram-honeypots like The Color Factory and Museum of Ice Cream, Sleep No More (an immersive theater experience), PingPod (play ping pong in urban areas, in 30 min increments), and my all-time-favorite, TeamLab exhibitions in Japan. Even professionally, we will prioritize exclusive conferences and random gatherings more than ever before. The value of non-scalable human-intensive experiences goes up. AI can’t disrupt any of these human experiences, but it can unleash them.
Creativity that is effective is creativity that moves us. Of course, all of these experiences are only as good as their quality and ingenuity…which brings us to creativity. It is the story and process behind an experience or piece of creative work that engages more than our eyes. So, i’d bet on the VALUE of human creativity going up in the age of AI. With such abundance of stuff catching our eye, we will crave story, value process, and be willing to pay for it. The prognosticators out there that suspect AI will somehow “replace” creativity don’t understand creativity. Sure, it will be easier than ever for all of us to make an image or a video. I love the fact that AI makes all of us more creatively confident. But people forget, these technologies also unlock the full potential of creators. 10x more cycles of discovery allow creative professionals to explore 10x more surface area of possibility which yields…10x better outcomes. Any designer knows: the more time (and patience…and interns) they have to explore more options, the better the results. The impact of creativity has always been a factor of two things: ingenuity and time. AI gives creative people with ingenuity the benefits of more time without spending time. Despite all of these breakthroughs in AI, one thing won’t change (and will become even more true and evident than ever before): The most effective creativity across art, culture, and marketing will be the creativity that moves us. This soul will continue to come from our humanity, and AI further unleashes this rather than replacing it.
In companies, humans will be redeployed towards higher-order tasks and non-scalable activities that make all the difference. With all this newfound human capacity liberated by AI, perhaps we will have MORE everyday unscalable experiences with the brands and services that permeate our lives. Manual on-boardings, why not? Personal shoppers at more brands for more customers? In-person training courses for enterprise SaaS? More brand building conferences and gatherings? We shouldn’t assume that liberated capacity means layoffs when it could be diverted to further differentiate a product and service - and drive customer success - in less scalable and finally economical ways. So, how will corporate education and leadership development change to accommodate this new world? We must focus on developing skill sets to help humans do things only humans can do. While executives have risen the ranks historically through productivity, they will increasingly do so via ingenuity and customer impact in ways that AI can’t deliver.
Finally, here’s a list of ideas and worthwhile mentions for those I work with and our smaller group of subscribers (and out of reach of the scrapers), including something I hope someone builds, called “Peanut Gallery.” Thanks again for following along…