The monied are all dogpiling money, water, electricity, pc hardware into it in their classic stupid tulipomania way (Just like with NFTs and Crypto before)

The management class in corporations are obsessed with it, they are delusional and think they can finally use it achieve their dreams of having no workers. They are forcing their workers to use it somehow. The are also foisting it unasked onto consumers in their products in the most stupid way. just wrecking their products.

And the reality of it is that.. The current crop of LLMs are excellent chatbots; they're very good at fooling a human into believing that they are intelligent.. They're good for shitposting and making silly images.. They have a few other quite situational niche uses, but that's about it.

I'm just pissed off at the stupidity and waste of it all. Like the Easter Islanders, our elites are driving our society straight into the ground.

▲

parasubvert 4 hours ago | parent [-]

Except this isn't like tulips or NFTs or Crypto; AI is actually useful (far beyond being a chatbot), and has real supply & demand, vs. a being pure speculative investment.

Yes, there's too much hype, and forcing the use of it does no one any good. And for sure a lot of that demand detached from results, and this means the demand will ebb. But I don't think this is like the dot-com bubble, it's more like the OPEC oil embargoing the 70s where soaring prices detached from results led to a collapse in demand that lasted a long while.

	▲	hannasanarion 4 hours ago \| parent [-]
		Why is it not like the dot-com bubble? The dot-com bubble is the most direct analogue. It is a genuinely useful technology that people are genuinely willing to pay lots of money for. Crypto and NFTs were speculative bubbles, the assets had no inherent value whatsoever, people bought exclusively because they hoped to sell to a bigger fool later on. Dot-com and AI are tech bubbles. The new technology is transforming a lot and growing fast and people are buying in in incredible numbers. But every technological adoption is an S curve, with an exponential phase followed by a logarithmic phase where it asymptotically levels off. The tech bubble forms during the exponential phase. As long as we are in the exponential phase, it is mathematically impossible to guess where the ceiling will be based on the trend alone, so any bet is justifiable. The crash comes when adoption inflects and growth slows and we all learn where the curve will level off. The losers in the bubble are the people who made bets on the ceiling being higher than where it actually lands, and people who made more conservative bets come out basically untouched. The outcome is dark fiber, or like we'll probably have in a few years, dark data centers. Infrastructure built by bankrupt firms who made big bets to get ready to service demand that is never going to come. But that infrastructure can be repurposed (the glut of bandwidth left by the Dark Fiber networks of 2001 basically enabled skype and voip).