“Doctor, it hurts when agents create unreviewable PRs.” “Don’t do that.”

I recently attended a talk, by an engineer at a large software company, on the topic of unreviewable PRs. The problem? When agents raise PRs with thousands of lines of LLM-written adds/deletes/edits, people can’t make sense of them. The solution? Throw more agents at the problem: reviewer agents that scan what coding agents have produced, identify problems, and triage them.

I don’t make software at industrial scale, so I can’t evaluate the claim that throughput gain justifies the absence of end-to-end human engagement. What I can say is that as I use Bram to bootstrap itself, I am fully engaged thanks to the workflow embodied in the tool.

Here’s the breakdown of languages in Bram.

Language

Lines of code