[okamiueru]

I appreciate your responses. It seems to me that your uses are to a much larger degree interpretative and also somewhat creative.

LLMs are great here. My concerns are mostly that its ability for factual accuracy is confused with its own expressed confidence. Assessment of one's owns abilities is usually optimistic in humans. LLMs just model that language. Facts are baked in as linguistic patterns, rather than a knowledge machine. And, you can often invert its answer to a scary degree.

There was recently an article on Forbes, where the journalistic "source" was the output of Bard.

The most terrifying part about LLMs to me, is not that it's sometimes wrong. But that it's often enough right and excels at certain stuff, that we use it as a source of knowledge. It really isn't.

On the other hand, the creative part, or dialogues regarding creative processes. It's mind bogglingly amazing.