Generation Got Cheap. Review Didn't.
The pitch for autonomous coding is a multiplier on how fast you write code. But writing was never the part a human has to stand behind, and the part they do stand behind got no cheaper, and probably got worse.
The pitch for autonomous coding is a multiplier on how fast you write code. But writing was never the part a human has to stand behind, and the part they do stand behind got no cheaper, and probably got worse.
Run git blame on a line an autonomous agent wrote and it points at a service account. The commit is spotless. The one thing debugging actually needs, a person who remembers why, already exited.
A public API can force you off an old version by making the endpoint go dark on a date. An internal platform has the same compatibility burden and no way to ever pull that lever, so nothing it ships ever gets retired.
Your audit trail proves the raw record wasn’t tampered with. It says nothing about which definition turned that record into the number you put in a deck, and definitions move.
When an agent generates a query, it binds to the column name, not the steward’s definition. A misleading name used to be a nuisance a new hire tripped over once. Now it’s a hazard that ships the wrong number every run.
An internal platform’s adoption chart looks like a vote and isn’t one. When the teams on it had nowhere else to go, the number everyone reports up means almost nothing, and the most captive consumer of all is the app you built it for.
Months ago I listed what a system needs before an agent can run unattended and treated the list as finished. It wasn’t. A stop button handles the runaway. It does nothing for the run that dies on step nine of fourteen.
Every governance program is built to add definitions and none of them can retire one. In a regulated shop that gets worse, because the instinct to keep every record gets misread as a ban on ever deprecating a definition. They’re different objects.
The moment you harvest a platform out of an app, the app stops being something you can change quietly. That standing cost lands hardest on the team that built it, and most extraction plans never budget for it.
The whole agent-reliability project is trying to engineer the one thing a tool can’t have: accountability. Data governance already named this. The agent is a custodian, and we keep trying to promote it to owner.