I'm currently obsessed with simulators / world models that can simulate world dynamics step by step. building this at lucid simulations, but I also dabble a lot in other kinds of generative models.
Previously worked on research in diffusion models with some brilliant friends. I also built a research chat over twitter known for its high taste, ml_chat.
questions that keep me up
why have we still not yet discovered a good interface for humans?
what interactions with modern chatbots like claude/4o will look like?
why is it so hard to analytically capture the "essence" of whats good looking?
experiments in progress
teaching ai to simulate using millions of observations
making diffusion models do weird things in real-time
hp ablations
people i keep tabs on
simo ryu - yes twitter is his actual blog
recent writing
loading...