John Schulman

John is currently a researcher at Anthropic, where he works on aligning large language models; some of his interests include scalable oversight and developing better written specifications of model behavior (like OpenAI’s Model Spec, Constitutional AI).

Previously, John was a cofounder of OpenAI, where he led the creation of ChatGPT, and from 2022-2024 I co-led the post-training team, which developed models for ChatGPT and the OpenAI API.

John received my PhD in Computer Science from UC Berkeley, and worked on robotics and reinforcement learning. Before that, he did a brief stint in neuroscience at Berkeley, and before that, he studied physics at Caltech.