so what the fuck is the market for humanoid robots

* useless in industry (or you'd see 'em long ago)
* sexbots? sure. i get the press release hype. a realdoll that talks (chatGPT) and moves
* waiting staff etc? basically novelties

BIG USE CASE: tech bubble hope

c'mon gimme actual application here

@davidgerard

I have good news/bad news!

There is a CLEAR demand for cheap-enough-to-be-affordable humanoid robots in care/nursing homes or other environments requiring hands-on labour interacting with human bodies or human-centric artefacts.

Bad news: cheap-enough approximates to "can compete with modern slavery wages" and the performance threshold is impossibly high—makes safe full-self-driving car software look like a fizzbuzz problem.

@cstross yep. they are frantically solving everything except the hard bit that is ~100% of the problem

@davidgerard @cstross Well, but look. If you see venture capital as the actual business they're in - selling dreams to vampires - then actually solving technical problems is an utter waste of resources and is in fact *counter*productive. A real solution can never be as sexy, and therefore as profitable, as a glorious future with a ground floor to get in at.

Thinking businesses should make things that work is as outmoded as thinking governments should foster broad-based prosperity.

@vivtek @cstross lol i am totally stealing that for my eventual writeup

(i can see this being >1000 words and the video taking four hours for post)

@davidgerard @vivtek The venture capital/startup model had a good 50 year run down Sand Hill Road in Palo Alto, modulo a couple of highly regrettable bubbles, but any system humans can invent that yields profit will eventually be gamed by other fraudsters, and I think large language model corruption caused by training using inputs based on garbage output is a great metaphor for what they're doing to themselves. VC is eating its own shit and saying "delicious!" in hope of attracting suckers.