Greg Brockman
Greg Brockman is co-founder and (former) President of OpenAI. Before OpenAI he was CTO of Stripe, which he joined as employee #3 in 2010 and helped scale from early startup to infrastructure for much of the internet's commerce. He dropped out of Harvard to join Stripe. At OpenAI he co-founded the organization with Sam Altman, Ilya Sutskever, and others in 2015, serving as CTO before transitioning to President. He took a sabbatical in 2024.
His two ingested essays are unusually candid reflections on what he actually did in both roles — not what a CTO is supposed to do, but what he discovered the job to be through trial and error.
The CTO discovery process
At Stripe, Brockman found himself unable to do all parts of the CTO job well. His response was systematic: he surveyed ~20 CTOs, hired a VP Engineering to cover his weak areas, and mapped which activities drained vs restored him. His key finding: most CTOs are facilitators, not chief architects. The "technical visionary" model is rare; most CTOs spend their time unblocking people, managing information flow, and making organizational decisions. See role-discovery.
The feedback loop problem: once you delegate deeply enough, you lose honest ground-level signal. Brockman's fix was returning to coding — not always the highest-leverage use of time, but the activity that recharged him and gave him direct feedback about what was actually happening in the codebase. See energy-management.
At OpenAI, a different pattern: find a problem bigger than any role (beneficial AI → find the right people → discover your unique contribution from what the org actually needs). His contribution was identifying engineering as the research bottleneck — projects like Gym and Universe existed because of this diagnosis. See role-discovery.