Ilya Sutskever ☆

Sutskever co-authored AlexNet under Hinton, drove OpenAI's scaling conviction as chief scientist, joined the board action against Altman and then left to found Safe Superintelligence — no products, one goal. His 2024–25 public remarks ('pre-training as we know it will end'; scaling's low-hanging fruit is gone) frame the field's central strategic debate.