In a recent Reddit AMA, Sam Altman admitted that OpenAI has been on the “wrong side of history” when it comes to its open-source strategy.
Despite previously favoring a proprietary development approach, Altman acknowledged the company’s need for a change in direction due to increasing competition, particularly from DeepSeek, a Chinese company that allegedly stole OpenAI’s IP.
Shift Toward Open Source
Altman expressed that OpenAI would likely adopt a more open-source strategy, including possibly open-sourcing older models that are no longer cutting-edge.
Kevin Weil, OpenAI’s chief product officer, also hinted that the company was considering revealing more of its models’ reasoning processes, unlike DeepSeek, which openly displays its models’ thought chains.
AI’s Role in Military Research
Amid discussions on OpenAI’s evolving strategy, Altman addressed concerns regarding AI’s potential use in destructive technologies, including nuclear weapons.
OpenAI recently partnered with the U.S. National Laboratories for nuclear defense research, with Weil expressing trust in the government’s responsible use of AI models.