Andrej Karpathy, one of OpenAI’s original 11 co-founders and former head of AI at Tesla, has joined Anthropic’s pretraining team. He will build and lead a new group focused on using Claude to accelerate pretraining research — effectively using AI to help train better AI. The hire is a significant talent win for Anthropic as it competes with OpenAI and Google for frontier model dominance.
Who Is Andrej Karpathy?
Karpathy is a Slovakian-born Canadian researcher who studied AI under Geoffrey Hinton and Fei-Fei Li. He was a founding member of OpenAI, contributing to early GPT model development before leaving in 2017 to lead Tesla’s Autopilot and Full Self-Driving programs. After five years at Tesla, he returned to OpenAI, then departed again in 2024 to launch Eureka Labs, an AI education startup.
What Will Karpathy Do at Anthropic?
Karpathy joins Anthropic’s pretraining unit, reporting to head of pretraining Nick Joseph. His mandate is to build a team focused on using Claude itself to accelerate the pretraining research process. This is a recursive approach — using AI to improve the systems that train the next generation of AI. The move signals Anthropic’s bet that research velocity, not just raw compute, will decide the next generation of frontier models.
Why Is This Hire Significant?
The hire is widely seen as Anthropic’s most high-profile talent acquisition. Anthropic was founded by former OpenAI employees who left over safety disagreements, and Karpathy is the most prominent OpenAI figure to join the rival lab. He joins other notable OpenAI alumni at Anthropic including Jan Leike and John Schulman (who later moved to Thinking Machines Lab). The hire lands the same week Anthropic is reportedly closing in on a $900 billion valuation.
How Does This Fit Into the Broader AI Talent War?
Karpathy’s move is part of a broader pattern of elite talent reshuffling. Anthropic has also recruited Ross Nordeen, a founding member of xAI, and Chris Rohlf, a Meta cybersecurity veteran with 20 years of experience. Meanwhile, Anthropic is reportedly in talks for a $30 billion funding round, and its revenue run rate topped $30 billion last month, surpassing OpenAI’s reported $24 billion.
Key Takeaways
- Andrej Karpathy joins Anthropic from OpenAI after founding both OpenAI and Eureka Labs
- Will lead a team using Claude to accelerate pretraining research
- Hire signals Anthropic’s strategy of AI-assisted research over pure compute scaling
- Karpathy is the highest-profile OpenAI figure to join Anthropic
- Move comes as Anthropic nears $900 billion valuation
- Anthropic projects $10.9 billion in Q2 revenue with first operating profit
Frequently Asked Questions
Why did Karpathy leave OpenAI? Karpathy hasn’t detailed specific reasons, but characterized his move as wanting to work on frontier LLM research during what he calls “especially formative” years.
Is Karpathy done with Eureka Labs? He stated he remains “deeply passionate about education” and plans to resume his work on Eureka Labs “in time,” but is now focused on Anthropic.
What does “using Claude to accelerate pretraining” mean? It means using Anthropic’s existing AI models to help design and optimize the training process for future models, potentially creating a compounding improvement cycle.