A Trace Proves What Ran, Not That the Answer Is Right
hugobowne · x · 2026-09-04
Hugo Browne opens his thread of 11 lessons learned from Hamel Husain about building AI products in the age of agents. The hook: an agent returned $4.21M of net revenue without showing the definition, tables, filters, joins, dates or fiscal calendar — 'what the hell is net revenue?' Point 2: a trace proves what ran, not that the answer is right; even after inspecting the SQL and every tool call, you still need a domain expert to catch the wrong table, definition or assumption.
Related event: Hamel Husain: Don't Build AI Products If You Won't Inspect Your Own Data(12 posts)→
More from coding & agent
- Clanker Cloud Opens Free Web Trial With $20 Worth of Starter Credits — tekbog · 2026-09-04
- Clanker Cloud Lets Anyone Build and Host Agents, With an Enterprise Sales Cautionary Tale — tekbog · 2026-09-04
- Running Grok bots like a company: AI project manager coordinates specialist agents — FinanceYF5 · 2026-09-04
- AI Finds Bugs Faster Than It Fixes Them: Engineers Grapple With CVE Backlogs — _jaydeepkarale · 2026-09-04
- Should Agents Govern Themselves? AAV Adds an External Action-Verifier Layer — CarlosMarreroAAV · 2026-09-04
- Dev spends $40 on classifier evals to cut costs: 'hard to use AI when you can't afford intelligence' — zeeg · 2026-09-04