Sébastien Bubeck
I work on AI at OpenAI. Prior to this I was VP AI and Distinguished Scientist at Microsoft, spending 10 years in Microsoft Research (first joining the Theory Group), and before that I spent 3 years as an assistant professor at Princeton University. In the first 15 years of my career I mostly worked on convex optimization, online algorithms and adversarial robustness in machine learning, and received several best paper awards for these works (STOC 2023, NeurIPS 2018 and 2021 best paper, ALT 2018 and 2023 best student paper in joint work with MSR interns, COLT 2016 best paper, and COLT 2009 best student paper). I am now more focused on understanding how intelligence emerges in large language models, and how to use this understanding to improve LLMs’ intelligence, possibly towards building AGI. We call our approach “Physics of AGI”, as we try to uncover at different scales (parameters, neurons, group of neurons, layers, data curriculum, …) how the parts of the system come together to create the striking and unexpected behavior of these models. A good starting point to learn more is this video and this one. Notable coverage:
|