Speaker

Greg Brockman

Greg Brockman

Co-founder and President of OpenAI, and its first Chief Technology Officer — the hands-on engineer-leader who assembled the founding research team and has driven the lab’s technical execution since 2015.

Before OpenAI, Brockman was the CTO of Stripe, which he joined early and helped scale. He left because Stripe’s mission, however important, was not the problem he wanted to spend his life on; making a difference to how AI plays out in the world was. In 2015 he and Sam Altman convened the dinners and the Napa offsite that became OpenAI, narrowing an initial team around Ilya Sutskever and setting the lab’s decade-long technical plan: solve reinforcement learning, solve unsupervised learning, and gradually learn harder things. He describes himself as an in-the-trenches leader who runs at problems from the front, and who builds environments in which others can do their best individual work.

Core positions

Brockman treats prediction as intimately connected to intelligence — predicting the next word out of Einstein’s mouth requires Einstein-level understanding — and holds that unsupervised pre-training and reinforcement learning are the same predictive technology over differently structured data. He champions iterative deployment: shipping intermediate systems so that the lab and society adapt to each other, rather than betting everything on one secret, untested launch. He argues the durable competitive advantage is not any single model but ‘the machine that makes the models’, which is why distillation and model theft matter less on an exponential. He frames safety as a core product feature and a matter of societal resilience — seat belts and roads for the engine, not the engine alone — and expects a compute-constrained world in which deciding where compute goes becomes society’s central question. Echoing Sutskever, he holds that building real value requires a willingness to suffer and to encounter reality as it is.

In the wiki