Kosta Derpanis (sabbatical in Zurich)

Your eyes don't see a whole scene at once — they dart around in a sequence of foveated glimpses. Modern vision models take in the whole image in one shot. That difference decides if a model can generalize to scenarios that are out-of-distribution. 🧵:

> replicate J-space on GLM 5.2 > train a reward model and run RL to reduce hallucinations > show me how this model makes cancer predictions Using our platform Silico is like having a team of AI researchers ready to run experiments like these. Private beta is open now. 🧵 (1/6)

Visiting Google Zurich today. My last visit, in 2018, I attended a public talk on something new called the Transformer. You could say it got everyone’s attention.