OpenAI co-founder Andrej Karpathy announces he's joining Anthropic

Andrej Karpathy, a 39-year-old Slovak-Canadian AI researcher and one of OpenAI’s original 11 co-founders, announced on May 19 that he is joining Anthropic. In a post on X, Karpathy stated he is excited to return to research and development at Anthropic, emphasizing the importance of the next few years in large language model (LLM) advancements. He also noted his continued commitment to education, planning to resume work in that area. Karpathy’s new role at Anthropic will involve building a team dedicated to using the company’s Claude model to accelerate pretraining research, as confirmed by Anthropic’s Head of Pretraining, Nicholas Joseph. Joseph, a former OpenAI employee, welcomed Karpathy to the team, calling him uniquely suited for the position. Karpathy’s career spans academic research, industry deployment, and education. He earned his PhD from Stanford under Fei-Fei Li, worked at Tesla from 2017 to 2022 as Director of AI, and later returned to OpenAI in 2023. His experience includes leading Tesla’s Autopilot computer vision team and contributing to synthetic data generation at OpenAI. Since leaving OpenAI in 2024, Karpathy has focused on AI education and open-source research. He launched Eureka Labs in July 2024, an AI-native school offering courses like LLM101n, which guides students in training AI systems. His public contributions include technical videos on LLMs and open-source projects like autoresearch. The announcement coincides with Google’s annual I/O developer conference in Mountain View, California, underscoring the high-stakes competition among leading AI firms. Karpathy’s move reflects Anthropic’s push to strengthen its pretraining capabilities, a critical area for advancing AI models.