Post #1090581

RE: https://toot.io/@bexelbie/116331613185287261 I kept trying to define "false positive" in code review and ended up somewhere different. When two humans debate code and nothing changes, both people still come out changed. When an LLM is wrong, nobody is changed. The worst part isn't the rate. It's being made to performatively justify yourself to something that won't learn from the exchange. https://www.bexelbie.com/2026/04/02/more-thoughts-sashiko-kernel