@judgedread

> There was a real business opportunity there, but most of the IPOs cashing in on it were scams.

Massive hype attracts sociopaths, and they tend to be good at adding hype.

> (whatever that is)

I've been putting on Hamming's lectures lately. (They were not properly ordered when I took the first pass, so I watched half of them and I don't know which half until I get ten or fifteen minutes in; not important.) Rewatched his lectures about AI, and one of his points was that nobody's got a satisfactory definition of "intelligence" or "thinking", and that, whether or not you believe in the possibility of a machine that is the equivalent of a human or you believe in a soul (he insisted that both beliefs will lead you to make big mistakes in your use of technology that will torch your career), the important thing to do is get comfortable with ambiguity, because it's going to get weird. I think he was overly optimistic about the timeline after which investors would jump back in (the lectures were from 1995, so not too long after the 80s AI crash), so it's nice to have run into it now. (He maintained that the question of whether a machine could do what a human does was pointless and that he was more interested in what humans using machines could do, and complained that this opinion was unpopular.)

> things like AI driven cars and trucks are going to be very profitable.

I'm skeptical that this autonomous vehicles will work out before the end of 32-bit timers or that this is the most profitable use for this sort of thing. I mean, look at Roomba: Amazon offered just $1.7B (about 1% of OpenAI's valuation) for the company that has an effective monopoly on these tiny devices that still have trouble navigating a house. Look at the stock: https://www.nasdaq.com/market-activity/stocks/irbt . Boston Dynamics sold the Hyundai for $1.1B five years ago, and both of those companies have plenty of fat government contracts.