OpenAI's Codex Product Lead: 'Steering, Not Rowing' as the New Model of Engineering Work
Summary
Tara Seshan (OpenAI product lead for Codex/ChatGPT Work) frames the shift as agents taking over execution while humans shift to opinionated direction-setting. Key distinction: coding is verifiable (tests pass/fail) so it automates fast, while most knowledge work isn't verifiable the same way — the process of talking to customers IS how you discover/trust a solution. Also: never fully automate 'writing-as-thinking,' only 'writing-as-reporting'; at OpenAI it's now 'mocks, not docs.'
Why this signal
A sharp, well-argued framework for where end-to-end AI acceleration does and doesn't generalize beyond code — directly useful for scoping which parts of Thomas Engineering's workflow (ideation vs. reporting) to push AI into next.
Original email
Lennys Newsletter · 8/30/2026
AI's third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI's product lead)