Anthropic Hires Former OpenAI Researcher Andrej Karpathy
Anthropic has hired Andrej Karpathy, a former founding member of OpenAI, to lead a team focused on using Claude to accelerate pre-training research. The move follows recent model releases and reported funding discussions at a valuation near $1 trillion.
thehindubusinessline.comAnthropic has hired Andrej Karpathy to its pre-training team. Karpathy announced the move on X on Tuesday. Karpathy wrote that the next few years at the frontier of large language models would be especially formative and that he was eager to return to research.
He began work this week. Karpathy was a founding member of OpenAI in 2015. He later ran AI at Tesla, returned to OpenAI in 2023, and left the company in 2024 to start Eureka Labs.
2025, Karpathy posted about a method he called vibe coding, in which users describe what they want and let the model complete the task. Collins Dictionary later named the phrase Word of the Year. In March, Karpathy described an experiment in which an AI coding agent ran unsupervised for two days.
The process produced 20 optimizations that reduced training time by 11 percent when applied to a larger model. He referred to the approach as autoresearch. According to Anthropic, Karpathy will start a team that uses Claude to speed up pre-training research.
The work focuses on large-scale training runs that build the model's core knowledge and capabilities. He will report to Nick Joseph. Karpathy previously ran a YouTube channel called badmephisto that taught competitive Rubik's Cube solving techniques.
