Anthropic engineers' new job: design tradeoffs — and telling Claude 'no'
ronbodkin · x · 2026-10-14
Part 3 of Anthropic's engineering thread: engineer responsibilities have shifted to supplying "understanding of the high-level inter-system components and the design-level tradeoffs," and "very often, the role of the engineers was to tell Claude 'no'" — human value now lies in judging which AI-generated changes shouldn't land.
More from coding & agent
- "Can't sleep, my agents require prompting": the agent-era insomnia — mallow610 · 2026-10-14
- Google's Gemini 4 launch marred by reports of hour-long hangs in Antigravity — op7418 · 2026-10-14
- Claude Code v2.1.299 auto-compacts 200K-token chats after 54 min idle — ashwin-ant · 2026-10-14
- AI notices who still reads every line: slopcore comic pokes at coding-agent complacency — TinfoilTricorn · 2026-10-14
- Sentry CTO: token burners are doing extreme low-ROI things, subscriptions are generous — zeeg · 2026-10-14
- Move rules from prose to deterministic constraints: poteto's token-saving agent vision — poteto · 2026-10-14