When code is nearly free: the real bottlenecks in scaling AI-built software
rseroter · x · 2026-11-07
- Pere Pages' 28-min article maps where teams slow down once agents write most code: deciding what to build, reviewing changes, and trusting tests.
- Key practices: write a short spec with examples that become tests; review by risk tier (money/login/PII/interfaces get full human review), not by habit; keep tests written from the spec separate from the code they check.
- Cheap fixes first: a shared agent instructions file, a merge-queue test gate, and a sandboxed agent workspace. Ship small, roll back automatically when errors rise.
More from coding & agent
- Managed Deep Agents can now be steered mid-turn while agents work — hwchase17 · 2026-11-07
- Every app should have an MCP server, argues Dan Vega — therealdanvega · 2026-11-07
- Google: stale TODO comments derail 1 in 7 coding agent sessions — rseroter · 2026-11-07
- Dev turns Runway AI video into a playable Unreal Engine game via agent — davidfromkansas · 2026-11-07
- Open Bot Avatar: MIT-licensed JSON spec for interactive AI agent avatars — kevinkern · 2026-11-07
- Developer rebuilds hospital notes as one document with a version-history timeline — BraydonDymm · 2026-11-07