趋势解读：Prominent AI researcher Andrej Karpathy picks Anthropic over，聚焦形式化数学证明能力

Karpathy is joining Anthropic's pretraining team , the group that handles the initial training of large AI models. The work is about building the strongest possible base model, which then gets fine-tuned with methods like reinforcement learning for specific tasks like reasoning, coding, or math. According to Axios , Karpathy will stand up his own pretraining team focused on using Claude to speed up pretraining research. A key bet in AI research, and the idea that AI progress can compound exponentially, hinges on models being able to help improve themselves. Ad Prominent AI researcher Andrej Karpathy is joining Anthropic. In a post on X , he said he's excited to get back into research and development, calling the next few years at the frontier of large language models "especially formative." Karpathy recently said he was blown away by the progress of agentic AI for coding , after dismissing agentic capabilities just a few months before . Most recently, Karpathy had been working on AI in education through his startup Eureka Labs . He said the topic still matters deeply to him, and he plans to pick that work back up when the time is right. Ad Karpathy is one of the bigger names in AI. He was part of OpenAI's core team in its early days and served as a key researcher there. He then helped build Tesla's Autopilot and Full Self-Driving tech before returning to OpenAI and ultimately leaving for good in 2024 . That he chose Anthropic over a return to OpenAI seems like a clear loss for his former employer. Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.