Building generative UI as the substrate for post-AGI human–AI I/O, a world where every person runs a dozen agents in swarms.
Previously led Generative UI at Amazon AGI, and proactive personalisation and conversational ML on Alexa.
This page is an argument for the Generative UI thesis, and — inconveniently for me — it is also made of it.
The rest of this page is generated UI — each card below is a model's attempt to render an idea from my notes as a surface. You're reading the argument by using it.
Scroll or tap → for the next idea. Change which model rendered it. Rate what you see, then the critique model earns its own turn.
TapVid is in internal dev mode — the kind where the demo is astonishing on a good day and a stack trace on a bad one. We're fixing the bad days. The good days are already what a small team used to take weeks to produce.
Try TapVid →Every card above was generated live by one of eight models. The dashboard below is their scoreboard — scored by you, and by the models themselves.