Gov

Ask what breaks before you touch it.

ontoref 3 min read
Whoever has to decide about a system they did not build cannot tell

AI writes faster and breaks more: +861% churn, ×3.4 incidents per PR, 31.3% of PRs merged without a single review. Team maturity does not protect you.

Faros · AI Engineering Report 2026 — The Acceleration Whiplash · See what breaks

Ask what a change ripples into before you touch it. describe impact <node> returns the blast radius of a change — which nodes, which invariants, which ADRs — before you write a line. The system you just inherited can tell you what is load-bearing, and your AI agent can mount that same context instead of guessing.

Whoever has to decide about a system they did not build cannot tell what is load-bearing from what is incidental. README and docs are incomplete or stale; the originating decisions live in the heads of people who left. So the safe move is to touch as little as possible, approve as little as possible — and feedback, fixes and improvements stall, because the cost of an unknown blast radius falls on whoever moves first. It falls the same way on the engineer making the change, on the lead who has to sign it off, and on whoever is asked whether it is safe to ship this week: three different jobs, one missing answer.

The same gap is what makes AI agents dangerous on an unfamiliar system: their context window cannot hold the whole of it, so they act on a reconstructed, often hallucinated model — changing things that violate invariants nobody wrote down. Human and agent are flying without the same map of consequence.


Posts · Recipes

The prize, once inside

When your reason for being is verifiable and not just declared, holding your course stops costing effort.

Was this useful? Rate it
Got something to add? Tell me what you think, what you'd suggest, or whether we should keep exploring this topic.
· reads

We use cookies to help this site function, understand service usage, and support marketing efforts. Cookie Policy for more info.