Andrej Karpathy, a 39-year-old Slovak-Canadian AI researcher and one of the first 11 founders of OpenAI, and former head of Tesla’s AI division, announced on Tuesday, May 19 that he is joining rival Anthropic.
As Karpathy wrote on his account on the social network X: "Personal update: I joined Anthropic. I think the next few years on the LLM frontier will be particularly constructive. I’m very excited to join the team here and get back to R&D. I always have a deep interest in education and plan to continue my career in it in the future."
Anthropic’s current Head of Pretraining, Nicholas Joseph, also a former OpenAI student, added more context to Karpathy’s new role at Anthropic in his post on X, writing: "I am happy to welcome Andrej to the Training team! He will be building a team focused on using Claude to accelerate the training research itself. I don’t think anyone is better suited to do it — I look forward to what we build together!"
An Anthropic spokesperson confirmed to VentureBeat via email that Karpathy will be starting a team focused on using Claude, Anthropic’s increasingly popular AI model, to accelerate self-training research. This will put Anthropic at the forefront of achieving the AI research goal of many around the world to develop "repeated self-improvement," that is, an AI capable of training its followers or improving itself incrementally with less, or ultimately no human intervention.
The announcement came on the same day as the start of Google’s AI-focused developer conference rival Google I/O at its headquarters in Mountain View, California, where new releases and announcements were expected.
History of Karpathy’s stories
Karpathy is widely recognized for bringing together three parts of the modern AI boom: academic research, large-scale corporate deployments and online education.
His personal website describes him as an AI researcher and teacher who was a founding member of OpenAI, later served as Director of AI at Tesla, and helped create Stanford’s first intensive course, CS231n.
OpenAI’s December 2015 launch announcement also listed Karpathy among the team’s founding members.
At Tesla, where he worked from 2017 to 2022, Karpathy led the Autopilot computer vision team and says his team handled internal data labeling, neural network training and deployment on Tesla’s custom chip.
He then returned to OpenAI from 2023 to 2024, when his website says he is building a team focused on generating artificial intelligence data and training — information that is directly related to Anthropic’s reported role in training.
Karpathy’s academic career began at Stanford, where he received his PhD under Fei-Fei Li and focused on neural networks in computer vision, natural language processing and the intersection of the two.
He also worked at Google Brain, Google Research and DeepMind, according to his website. His education includes an MSc from the University of British Columbia and a BSc from the University of Toronto, where he double majored in computer science and physics.
What will happen to Karpathy’s open source research and commitment to AI education?
Since leaving OpenAI in 2024, Karpathy has become one of the most visible AI community educators, publishing videos for technical and general audiences on large-scale language models and neural networks.
He also launched Eureka Labs in July 2024 as an “AI-native” school; its first product, LLM101n, is described as an undergraduate-level course that guides students through training their own AI system.
Working solo as a free agent for the past two years, Karpathy has also helped advance open source AI research through products and standards including autoresearch, an LLM-driven automated researcher that can run multiple hypothesis and test simultaneously, and LLM Knowledge Base, an independent memory and context storage system designed for AI to grow from such a library.
The big question is what happens to these and Karpathy’s open source AI efforts in general as he joins Anthropic, a lab that has championed open source by introducing its Model Context Protocol (MCP) technology standard, but is also famous for primarily deploying AI models and harnesses (like Claude and Claude Code).
Based on the last statement in his announcement post on X – "I always have a deep interest in education and plan to continue my career in it in the future" – it seems that at least his contributions to the native AI school effort will be put on hold as he studies Anthropic.
