There's a thing in Classics studies where you'll read surviving descriptions of Ancient Greek automata which attribute all manner of near-lifelike behaviour to them – then you look at the reconstructed plans for the automaton in question, and it's a device with roughly the sophistication of a wind-up mouse.
The broad consensus seems to be that the authors of such descriptions are exaggerating for clout. For my part, I look at all the people in the year 2026 who've managed to genuinely convince themselves that LLMs are not only sapient, but smarter than they are, and I think: hmm.
