Andrej Karpathy, a co-founder of OpenAI and former director of AI at Tesla $TSLA, has joined Anthropic, he announced on Tuesday.
Karpathy will build a team focused on using Claude to accelerate pretraining research at the OpenAI rival

San Francisco Chronicle/Hearst Newspapers / Getty Images
Andrej Karpathy, a co-founder of OpenAI and former director of AI at Tesla $TSLA, has joined Anthropic, he announced on Tuesday.
"I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative," Karpathy wrote on X $TWTR. "I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time."
Karpathy started this week on Anthropic's pretraining team, which handles the large-scale training runs that give the company's Claude models their core knowledge and capabilities, according to TechCrunch. The team is led by Nicholas Joseph, another former OpenAI employee and an early Anthropic hire. Anthropic said Karpathy will build out a group focused on using Claude to accelerate pretraining research itself.
Joseph welcomed the hire on X. "Excited to welcome Andrej to the Pretraining team! He'll be building a team focused on using Claude to accelerate pretraining research itself. I can't think of anyone better suited to do it," he wrote.
Karpathy was a founding research scientist at OpenAI, where he focused on deep learning and computer vision before leaving in 2017 to join Tesla. At Tesla, he led the Full Self-Driving and Autopilot programs before departing in 2022. He returned to OpenAI for about a year before leaving again in early 2024 to start Eureka Labs, a startup applying AI to education. He also created an online course called Neural Networks: Zero to Hero and posts lectures on AI to a YouTube channel, according to TechCrunch. He also coined the term "vibe coding," according to Business Insider.
The move places Karpathy at one of the most compute-intensive and expensive phases of building a frontier AI model. Pretraining is where models acquire foundational knowledge before any fine-tuning or instruction training takes place.
Anthropic has been in discussions on a $30 billion fundraising round that would value the company at $900 billion, surpassing rival OpenAI's most recent valuation of $852 billion. AI coding products, chief among them Claude Code, have accounted for much of the company's rapid revenue expansion.
Join 500,000+ readers who start their day with Quartz.
By subscribing, you agree to our Terms of Service and Privacy Policy.