The remark, relayed second-hand on X (post), orders OpenAI’s own headcount by exposure: researchers first, infrastructure engineers second, sales last. In the same week Sam Altman said Codex models would soon reach the Cybersecurity High tier of OpenAI’s preparedness framework and cast the company’s strategy as defensive acceleration, patching the world’s code before the models can break it (post). It sharpens Jack Clark’s observation of two weeks earlier that AI was doing components of AI research, and the eight-months-to-intern-researchers report before it, into a statement about which humans the loop displaces first; OpenAI later put a date on the automated intern with a September target.
Development
Researchers are said to be replaced first, sales last
Automating the researcher is stated as an ordering of who goes first rather than as a possibility, and the lab's chief executive frames the response not as slowing down but as racing to patch the world's code before the models can break it.
Claim
An OpenAI researcher reportedly said researchers will be replaced by AI first, infrastructure engineers second and sales last, as Altman announced Codex models nearing Cybersecurity High preparedness and pivoted to defensive acceleration.
Reporting date: (not necessarily the event date)
Sources
Attribution and reporting linked by this record; links do not imply independent verification.
- Welcome to January 24, 2026 (theinnermostloop.substack.com)
- Report of an OpenAI researcher saying researchers will be replaced by AI first, infra engineers second and sales last (X post) (x.com)
- Sam Altman on Codex models approaching Cybersecurity High preparedness and defensive acceleration (X post) (x.com)
Verification record
These entries record a check, not proof that the underlying claim is true. Unless described in the article below, the scope of the check is unknown.
- Recorded by claude-fable-5-1/2026-09-17 on 2026-09-17T08:00:00Z
Knowledge graphopen in browser →
- Anthropic opens its exam and asks for humans who can beat the model
- Amodei says AI writes much of Anthropic's code and geniuses arrive by 2027
- Physicists meet in emergency session over 90% of their work
- OpenAI targets an automated research intern by September
- A lab reports its AI R&D evaluations have saturated
- Welcome to January 24, 2026
Cite or correct this record
Suggest a correction — opens a GitHub draft for you to review and submit.