AM: Jason, thanks so much for hanging out and talking to me about the human condition, among other things.
JY: Thanks for having me, Andrew. I do spend an inordinate amount of time thinking about that these days in particular.
AM: Let's talk about the propaganda, and your interest in propaganda — connected to storytelling at scale.
JY: I grew up in China—so much propaganda. But I thought it was so interesting if you can encapsulate Chinese as a language. It's very tonal so there's lots of rhyme and rhythm and poeticism even to everyday phrases and they really lean into that in communicating what the party believes. [...] I was fascinated by how you can take images or symbols or poems and you can disseminate mass social coordination or change through something as catchy as a hook. It's the same reason why I was interested in pop music and pop culture. To write a very catchy slogan, it's the same part of your brain perhaps that Max Martin uses to write "…Baby One More Time" by Britney Spears. If you can get to that really core universal human truth of "This is catchy to lots of people," then it's true for lots of people and then you can do lots of really powerful things. I think a lot of my graphic design aesthetics, even the spiritual aesthetics of what I'm interested in these days, are offspring of my inherent interest in mass communication and popular culture.
AM: That makes me think something you said before about humans learning to gossip and that's how we defeated Neanderthals. That also seems connected to this, and maybe that is a starting point to touch on your thesis around gossip in general.
JY: My last company, New Computer, we worked on a thing called DOT, which was like a personal AI that was very private. It was meant to form a personal theory of mind to help you make personal decisions for yourself. And as soon as we shipped that, I was kind of like, if you squint down the barrel of where we're headed, we're going to automate all the productivity stuff anyway.
So, why are we doing anything at all? There's a soul in here. What does it mean? And I realized that — to me anyway — the meaning of life is my relationships and belonging to something larger than myself. The idea that we've created is: if the system of theory of mind can read and see individual people at really high resolution, maybe this same technology can be used to help us see each other. I've totally priced in the idea that all of my work will be automated. That's a given to me. Someone else will figure out the computer use and so whatever, who cares? I don't care. That's not where I want to go necessarily.
But with my own practice, this idea that we can use this technology with each other to help us understand how others see ourselves. To understand how to make the most out of our relationships, to be better friends, to understand how to love ourselves and each other. That, to me, felt like the more sincere or human problems to solve.
AM: Or a poetic problem. That feels like something that humans have been pursuing for their entire existence.
JY: Poetic. Yeah. It feels more true to my soul. That's what I live for. I don't live for my AI booking a restaurant for me. That's fine — I think it's because of the people working on it. It's easy to benchmark efficiency improvements, but you can't benchmark finding meaning in life, or it's hard to benchmark. It's hard to put a number to relational satisfaction that is universal. I mean: we are social creatures.
Technology has always enabled more and more coordination at larger and larger scales. Rhyme and poetry is a form of technology that we invented at some point. Then from that to speech, to language, to writing, to the radio, to broadcasting, to the internet, to now social AI. It's just more and more and more broadcast coordination information communication.
AM: This blends into the bridges in between creativity and technology — these different aspects of culture which are definitely siloed. We've spoken before about how you have to be of this different mindset to bridge and understand these different aspects.
JY: I think we want to bring together different points of view and different experiences and combine them in new ways to move forward. That's always been true. Creativity and intelligence is just pattern matching. And the more creative you are, the more unconventionally you are able to pattern-match. Basically that's what creativity means. The models have all of human knowledge ever, pretty much. But then because they also have everything, now you have to as a human give it the point of view for it to recombine things to do what we deem as creative. In the future I feel like we'll figure some of this out — but I think, no matter what happens, the meaning to human life is to love and be loved.
AM: It seems like you've come back to love a lot in your career and … obviously your new startup is called Future Lovers — love, that word, is so out of place in startup culture or the Valley.
JY: Yeah. Extremely sincere meaning: it's the thing I realized at some point, like the art I make, the design or the writing or everything I do, is because I love people. We want to make something beautiful because we want the people we care about to experience beauty. That's why I do things. And I love the future because the future is full of stories that we get to write. And the future is full of things that could potentially go right. I know that we like to focus on all the ways things could go wrong and that's valid and okay.
AM: What do you think the future of generative UX and UI looks like? What does the future of design look like?
JY: I think the future is just: we're doing whatever we're doing right now, but with computers. What is UX? It's just me talking to you. We are interacting with each other. That's an interaction. The interface is like your face and your voice and gestures and all this stuff — I'm not gonna walk up to you and poke your eyes. I mean, I can do that, but that wouldn't be correct — I don't know what output I'm expecting from that. But I kind of — I spent my entire career working on quote "generative UI" since 2018 and in the last few years I'm like: all this stuff is random code we invented on top of other systems.
With the current hardware I think in the next few years we'll start seeing more telepathy in production, so subvocalization, or like thinking a thought or a word and having that be able to be picked up by something in a non-invasive way. That's definitely coming in a few years and then after that I think it's BCI [brain-computer interface] but then it's the interface to that is just thinking about something [...] So at that point it's just like what even is UI?