John Schulman, who co-founded OpenAI and has been credited as one of ChatGPT's leading architects, leaves Anthropic, after joining the startup in August 2024
Jon Victor / The Information :
Context & Ripple Effects
Schulman’s move to Anthropic was framed as a shift toward alignment work after his OpenAI departure, linking a ChatGPT architect to a rival founded by former OpenAI employees and built around Claude. His August 2024 move to Anthropic therefore carried significance beyond a routine executive hire.
His exit closes that short stint and precedes reporting that he would join Mira Murati’s startup. The reported next move to Murati’s venture underscores how a small pool of experienced frontier-model builders is circulating among competing labs and new entrants.
First-order effects
- Anthropic loses a prominent researcher and OpenAI co-founder soon after hiring him, while Schulman becomes available to another AI venture.
- The departure shifts attention from Anthropic’s ability to recruit established frontier-model talent to the retention of that talent once hired.
Second-order effects
- Rival labs and newly formed startups gain a sharper recruiting case for senior researchers with rare experience building widely used AI assistants.
- Anthropic and peers may need to compete not only on research missions such as alignment, but also on the autonomy and scope offered to senior technical leaders.
Third-order effects
- If short tenures among prominent AI researchers persist, frontier-lab advantage may depend less on attracting individual star hires than on building organizations that retain cohesive research teams.
- The movement of founders and senior model builders toward new ventures could broaden the field of well-funded AI labs, even as the underlying talent pool remains concentrated.
The trend: Frontier AI is becoming a contest for scarce technical leadership, with experienced researchers increasingly able to reshape competition by moving between incumbent labs and new startups.