Back to the Frontline After a Year and a Half! Karpathy Joins Anthropic, Previously Hinted at Anti-Drift Mechanism in March

According to Dynamic Beating monitoring, on May 19, former Tesla Autopilot Director and OpenAI Co-founder Andrej Karpathy officially announced on the X platform that he has joined the cutting-edge AI lab Anthropic to return to the forefront of research. This move comes nearly a year and a half after his departure from OpenAI in February last year.

This highly anticipated return was not without warning, as its deep motivation had been publicly revealed two months ago. In a podcast interview titled "No Priors" released on March 20, Karpathy openly expressed his unease about being detached from the forefront lab. He highlighted that the black box barrier constructed by closed-source tech giants is causing inevitable drift in external researchers' technical intuitions and judgments.

At that time, he disclosed his extreme wariness of being completely controlled by a single closed-source entity and his inherent skepticism of centralized power. For a simple reason, he never wanted global intelligence to be monopolized by two or three closed-door giants. Therefore, in the podcast, Karpathy envisioned an ideal "secondment" status, where by going to the forefront lab for a period of research, he could both stay connected to the latest technology and not entirely lose his independence. This formal joining of Anthropic clearly fulfills this secondment concept to prevent drift.