Reflections on my time at OpenAI

2 months ago, I resigned from OpenAI. In total, I worked there for 16 months. I chose to leave because I believe the labs will reach AGI soon and it is time to start thinking about what other problems need to be solved. Overall, I’m incredibly thankful for my time at OpenAI, to have had the chance to contribute to AGI, and I really couldn’t have asked for a better team to work with. Some quick thoughts on my time at OpenAI:

It is incredible how quickly AI has progressed. When I joined in March 2025, o3 and codex had not been released. I had to write code by hand, get destroyed in code review, and dig through the intricacies of infra. Thanks to codex, most researchers do not need to do this anymore. In a little over a year, we went from IMO gold to solving millennium prize problems. Things that were once hypotheticals on LessWrong actually happened. It’s been an honor to play a part in this.

With the growth of the company, the last year has also spurred a lot of personal career growth. I found OpenAI to be a place very willing to give people opportunity and judge based on quality of work, not seniority or experience. I found myself doing things at OpenAI I thought would take me many more years to achieve, and I’ve been pushed to believe in myself, be more ambitious, and make the things I care about happen. Of course, there are so many more things to learn, and I’m very thankful for all the mentorship and support I got at OpenAI.

At the crux of it, time feels so valuable. It is crazy to think about how I could have spent the last year taking classes at Stanford instead of pushing the frontiers of intelligence. So many humans have lived and died, and I feel so lucky to be alive at this moment of history and be part of it.