Former OpenAI Researcher Joins Anthropic, AI Talent War Heats Up

On local time May 19th, Andrej Karpathy announced on X that he has joined Anthropic, expressing his "extreme excitement to return to a research and development position," and stating that "the next few years at the frontier of large language models will be particularly formative." Anthropic confirmed that Karpathy has joined the pre-training team responsible for large-scale model training, and the work of this team directly determines the core knowledge and capabilities of Claude.

Karpathy's resume carries significant weight in the AI circle. He joined OpenAI in 2016 and moved to Tesla in 2017 to lead the development of the Autopilot self-driving project. Musk referred to him as "the second person in the field of computer vision globally" in an internal email from 2017, which was later exposed during the trial of Musk's lawsuit against OpenAI. After leaving Tesla, Karpathy founded AI education company Eureka Labs, and this time joining Anthropic indicates his return to cutting-edge research.

The context of this move is intriguing. Less than two weeks ago, Ross Nordeen, a former senior executive of Musk's xAI, also announced that he has joined Anthropic and is currently responsible for building and expanding Claude's infrastructure. Both talent shifts occurred after Musk's lawsuit against OpenAI and CEO Sam Altman ended in defeat for Musk, and outside parties generally interpret this as a sign that the talent battle between the three companies is heating up.

At present, both Anthropic and OpenAI are privately-owned companies valued at around $1 trillion. They form direct competition in AI models: OpenAI has an advantage in the consumer market with ChatGPT, while Anthropic is more competitive in the field of enterprise applications. Karpathy's joining will directly strengthen Anthropic's capabilities in fundamental model development.