A New Perspective on Adversarial Perturbations · Simons Institute / YouTube
Known for
Left OpenAI May 2026; building on AI economics.
Why chosen
He links rigorous ML theory, AI safety and the economics of intelligence. As an MIT professor his adversarial-robustness research reshaped how the field thinks about reliable models. He spent almost three years at OpenAI, where he led the Preparedness team and later worked on reasoning. In May 2026 he left to invest in and build projects on AI's economic impact. Robust, then and now.
A New Perspective on Adversarial Perturbations · Simons Institute / YouTube
Role
Former Head of Preparedness, OpenAI (former)
Status
New ventureLeft OpenAI in May 2026 to invest in and build projects on AI's economic impact.