I don't voluntarily use #LLM #AI systems out of principles, though I don't judge anyone for using them.

Still, I keep informed through people using it and people researching their effects. And so far I've seen a pretty massive discrepancy between the favorable impression these tools make on individuals while on aggregate, the value of the output is pretty low. That's why I dub them fooling machines: Even when they aren't intended to manipulate people, the natural language interface still makes a good impression, regardless of the actual quality of the output.