Sam Altman just went on record about Elon Musk's time at OpenAI, and it wasn't pretty. In testimony for Musk's lawsuit against the company, Altman said Musk did "huge damage" to OpenAI's culture during his early involvement.
The specific issue? Musk pushed OpenAI president Greg Brockman and former chief scientist Ilya Sutskever to rank researchers by performance and "take a chainsaw through a bunch." That's classic Musk management, the kind of forced ranking system he's used at Tesla and later at Twitter.
Altman acknowledged this was Musk's known style but said it fundamentally didn't work for OpenAI. "I don't think Mr. Musk understood how to run a good research lab," he testified when asked about morale after Musk's departure.
This matters because it reveals the tension between startup velocity culture and research lab culture. AI labs need to retain top talent and foster collaboration, not create internal competition. The stack ranking approach might work for scaling manufacturing or shipping product updates, but research requires a different environment.
The testimony is part of Musk's ongoing legal battle with OpenAI, where he's claimed the company abandoned its original mission. But these details suggest the cultural clash ran deeper than just disagreements about commercialization.
For anyone building or working in AI teams, this is a reminder that management style isn't one size fits all. What works in a move fast and break things environment can break the kind of patient, collaborative work that produces breakthrough research.
