Vercel CEO Guillermo Rauch: His Agent Runs Product Reviews, He Only Joins When It Fails
cen6wkf · reddit · 2026-09-07
Vercel CEO Guillermo Rauch says he hasn't sat through a product review meeting in months — his own agent runs them, and he only shows up after it fails. His rule: if he DMs you, his agent already tried and failed first.
The post argues this personal habit is becoming enterprise policy: Google Cloud published nearly identical governance advice the same week — build automatic escalation points so agents act autonomously and humans are looped in only on genuine failure.
Key takeaways:
- The McDonald's kiosk analogy: automation is gradual but irreversible — what can be automated, will be
- The jobs being replaced now are unwritten coordination work — syncing, pestering, "let me check on that" — not hard technical work
- Winners design the replacement rather than defend the coordination role
- Rauch laid out the executor/verifier split of engineering roles back in June, consistent with this admission
Source: a clip from the Knuckle Up with Nakul podcast.
More from AGI Musings
- NBER: Only 2.9% of US consumers pay for AI; retention as low as 20%, pricing power at risk — avicgoldfarb · 2026-10-26
- AI researchers guessed robots wouldn't outrun humans until 2034 — it already happened — ben_j_todd · 2026-10-26
- ESR on the Bugpocalypse: locking down frontier models only hands the game to Chinese open weights — colinmcnamara · 2026-10-26
- AI folks, stop saying "biology" when you mean drug development — KevinKaichuang · 2026-10-26
- Anthropic: junior tech hires fell from 27% to 9% as agent hours hit 6x employee hours — thione · 2026-10-26
- Anthropic's Jascha: you may have only 1–5 years of intellectual contribution before AGI — GabGarrett · 2026-10-26