OpenAI's quiet co-founder steps out
Wojciech Zaremba makes the case for the OpenAI Foundation's $25 billion "AI resilience" bet. Also: Anthropic readies for IPO, Altman goes to Michigan, and more.
Happy Monday. One of these days, I’ll send the newsletter earlier in the day, I promise. I’m in New York City later this week for some meetings and the new media soirée I’m co-hosting with a16z. Today, I’ve got an exclusive chat with OpenAI co-founder Wojciech Zaremba.
Wojciech Zaremba recently bought a copy of “House on Fire,” a 2011 memoir by epidemiologist William Foege about the campaign that wiped smallpox off the planet. He’s using it as a guidebook for executing what is about to become one of the largest philanthropic efforts of all time.
Zaremba is one of OpenAI’s least well-known co-founders. He has spent more than a decade at the company across a range of efforts, from leading its early robotics efforts to starting the team that guides OpenAI’s personality and what became reasoning models. In March, he left the frontier research world to run AI “resilience” at OpenAI’s nonprofit foundation.
Zaremba and I spoke ahead of a post that the OpenAI Foundation published Monday morning titled “Resilience in the Age of AI,” which names four areas it will initially fund: biosecurity, cybersecurity, model safety, and AI’s effect on kids. After $100 million for fighting Alzheimer’s with AI in April and $250 million for researching “economic futures” last week, the initial $25 billion grant machine Zaremba helps oversee is spinning up.
“Even if OpenAI makes the safest models out there, that doesn’t solve the problem of jobs [going away],” Zaremba told me during an interview. His use of the word “resilience” is telling; it implies that, even with fear of AI at an all-time high, he and OpenAI are comfortable publicly framing AI’s advancement as inevitable and something to be withstood. “There is no way to align fire. There is no way to align electricity,” he said. “We need to be resilient to the outcomes.”
