OpenAI co-founder Andrej Karpathy joins rival Anthropic for pre-training research
Another prominent defector from OpenAI lands at Anthropic to build frontier models.
Andrej Karpathy, a founding member of OpenAI and former director of AI at Tesla, has joined Anthropic, the AI lab founded by ex-OpenAI employees. Announcing the move on X, Karpathy said he is "very excited to get back to R&D" and that "the next few years at the frontier of LLMs will be especially formative." According to Axios, Karpathy will join Anthropic's pre-training lab and lead a new team focused on using the Claude model to accelerate pre-training research.
Karpathy has had an on-again, off-again relationship with OpenAI, leaving in 2017 to lead AI at Tesla, returning in 2023 for about a year, and departing again after the brief ouster of Sam Altman. His move comes amid an intensifying cold war between OpenAI and Anthropic, with the latter positioning itself as the more ethical alternative. Karpathy, known for coining "vibe coding" and "tokenmaxxing," brings both serious technical credibility and a flair for AI hype. The hire signals that Anthropic is winning the talent war, even as its rivals struggle with internal drama.
- Former OpenAI co-founder Andrej Karpathy joins Anthropic to lead a new pre-training team using Claude.
- Karpathy had a long, fractured history with OpenAI, leaving twice over eight years.
- The move reinforces Anthropic as the top destination for AI talent, especially former OpenAI employees.
Why It Matters
High-profile talent flowing to Anthropic signals a shift in the AI industry's center of gravity away from OpenAI.