OpenAI co-founder Andrej Karpathy joins Anthropic to work on pre-training

In short: Andrej Karpathy, a co-founder of OpenAI and former AI leader at Tesla, has joined Anthropic to work on “pre-training” for its Claude AI models.

What happened

Andrej Karpathy said in a post on X that he has joined Anthropic. He wrote that he is excited to return to research and development and that he plans to continue his interest in education over time.

Anthropic told TechCrunch that Karpathy started this week. He is working on pre-training under team lead Nick Joseph. Pre-training is the big, early training step where an AI model learns general knowledge by studying huge amounts of text (like sending a student to read a very large library before starting a specific job).

Anthropic also said Karpathy will start a team focused on using Claude to speed up pre-training research. The company described pre-training as one of the most expensive and computer-heavy parts of building top AI models.

Karpathy is well known in the AI field. He previously worked at OpenAI, left in 2017 to join Tesla, and led Tesla’s Autopilot and Full Self-Driving efforts before leaving in 2022. He returned to OpenAI for a year, then left again in 2024 to start Eureka Labs, a startup focused on education. It is not clear if he will continue working on that startup.

Why it matters

For regular people, this is a sign that Anthropic is investing in improving how Claude is built at the earliest stage. That can affect how helpful, accurate, and reliable future AI assistants feel in everyday use.

Source: TechCrunch AI