I don't deny they're incredibly useful - I use them regularly myself for help with everything from the Wolfram Language to complex SQL queries.

I think that the point I'm trying to get across is that they're great at what they do, if they can be verified to be correct, but insofar as offering a new wonder in communication or a path to AGI I think that hype has very much overtaken reality.

They are a better interface to search, or address a personal assistant, but should not replace deterministic, factual results with their own imagination.