Is anyone else tired of human in the loop?

One of the biggest issues with AI is the slop that it tends to produce as "Documentation" once you have read one you end up having read them all with "varying" degrees of nuance. So we have ended up asking AI to produce succinct and plain english documentation...ya know if you can explain it simply then you understand what you are talking about 🤣

Some related terms that have become popular: Orchestrator, Evaluator and Judge. When humans orchestrate they have control and responsibility. When humans evaluate they have criteria and context. When humans judge they have authority and accountability. Maybe there's something in making these role types specific. One person or a group of people could be all three of course. Yet perhaps that's what human in the loop is trying to infer instead of a check box exercise. Like any new term, it's all open to interpretation and a multitude of usage. I'm with ya, 'human in the loop' feels like it's becoming a catch all to say "Don't hand it all to those AI agents, make sure to check in at some point in your workflow to see how they are getting on with the extremely important and risky work you've handed off to them... looks good to me!" 🫠

It's a good observation and one I hadn't really considered much. I still baulk at the term "hallucination" which rather obfuscates the fact the output is inaccurate at best and just plain wrong at worst.