VC shares the 3 AI questions he now asks in tech due diligence: logic location, permissions, cost per request
julsimon · x · 2026-10-01
An investor shares the AI questions he now adds to tech due diligence, arguing that traditional vendor DD verifies code quality and team competence but not whether the company can actually build AI products on its platform.
- Where does business logic live? If it sits in the screens, every AI feature requires rebuilding the product—price that rebuild.
- Check permissions. If "edit" covers everything, you can't safely hand anything to an agent; he calls this the most common blocker.
- Cost per request. Flat pricing on a feature whose cost scales with usage is a margin problem that grows with adoption.
Every answer must be demonstrable on screen, in code, or on an invoice—"we plan to" is a no.
More from coding & agent
- CMU releases CUA-SWE, a benchmark uniting computer-use agents with visual software engineering — CarnegieMellonU · 2026-10-01
- Médula open-sources data on coordinating parallel Claude Code agents: Jev decider escalates 61% of real writes — jokiruiz · 2026-10-01
- Three agents worked the same account blind — one team's fix with a shared memory layer — Davnys · 2026-10-01
- Skip detailed feedback: coding agents fix their work with just "this sucks, improve it" — bosmeny · 2026-10-01
- Physics YouTuber shows Claude fully automating his explainer videos, 3.5 years after asking when AI could — AndyMasley · 2026-10-01
- GitHub Unveils Agentic Engineering System: A Framework for Scaling AI Agents — marlene_zw · 2026-10-01