"People who think they're immune to marketing are the biggest suckers of all" always been bad, but it's worse than ever because a bunch of people have decided that the most reliable way to judge AI performance is its vibes on X The Everything App. I can't even really fault places like doomers.ai when their marks are practically asking to get conned.
I appreciate you putting words to the astro-turfing. two areas llms have helped me are in (i) giving me work and (ii) learning how people in my life repeat what they hear from “knowledgeable” people. And not saying the voices proselytizing llms aren’t knowledgeable but I bet there’s an agenda behind their words (ie “i have something powerful” and “give me money!)
A functional difference is the probabilities associated with each option. True, you could use the constrained next-token distribution at the appropriate generation step, but those would not be calibrated by any means, which Jev's claim to be. An empirical analysis would be interesting.
Obscene marketing is all you need it seems.