Andrej Karpathy Joins Anthropic Pretraining Team
Andrej Karpathy announced Tuesday that he joined Anthropic and will work on the company's pretraining team. The move follows his departure from OpenAI in February 2024 and the launch of his education startup Eureka Labs.
Andrej Karpathy announced on Tuesday that he has joined Anthropic and will work on the AI company's pretraining team. Karpathy, a founding member of OpenAI and former director of AI at Tesla, stated that he is excited to return to research and development work. He added that he remains passionate about education and plans to resume related projects in the future.
Karpathy helped launch OpenAI as a founding research scientist. He later served as Tesla's AI director before rejoining OpenAI in 2023 and leaving the company again in February 2024. After departing OpenAI, Karpathy founded Eureka Labs, an AI-focused education company.
Anthropic said Karpathy will work on large-scale testing of its Claude models under the pretraining team led by Nick Joseph, a former OpenAI employee who joined Anthropic early. Business Insider reported that interest in Anthropic has increased in 2026 following advancements in its Claude Code and Cowork tools and a dispute with the Trump administration.
The same report noted that Anthropic's rivalry with OpenAI has intensified, with OpenAI CEO Sam Altman recently accusing Anthropic of contributing to hostility directed at him.
Transparency
Reported by a single outlet. This score reflects source tier and factual specificity — corroboration is limited with one source.
Story details
Related Stories
abcnews.go.comTrump Signs Executive Order Prioritizing AI for Cybersecurity Innovation
President Donald J. Trump signed an executive order on June 2 directing federal agencies to accelerate artificial intelligence development for protecting critical infrastructure. The order reverses earlier emphasis on slower deployment and risk reviews.
nbcnews.comTrump Signs AI Executive Order Promoting Innovation While Requiring Security Reviews
The order directs federal agencies to promote advanced AI development while addressing security concerns and reduces government review compared with an earlier draft.
nbcnews.comTrump Issues Executive Order for Voluntary 30-Day AI Model Sharing to Boost Innovation and Cybersecurity
President Trump signed an executive order on artificial intelligence and cybersecurity on June 2, 2026, establishing a voluntary 30-day pre-release window for frontier models and an industry collaboration on vulnerability scanning.