Really don't understand why it's so taboo to ask AI people how they're testing their automations for failures, cost, etc... these are normal questions any business worth its salt would ask for any other tool.

Guess people don't want to draw attention to the fact that they're expensive, constantly put out bad data (hallucinations, outdated info, poisoned content, etc.) and just aren't worth the investment (monetarily or time) but I get it. Some people really think they're getting in on this early and don't believe that tool usage has moral implications.