Kat S on Nostr: The dot-com bubble was a gold-rush, driven by the greed of founders and investors. It ...

The dot-com bubble was a gold-rush, driven by the greed of founders and investors.
It wasn't met with the same kind of opposition, partly because fewer people understood the implications (the internet had really only just reached widespread public awareness) but also partly because it was basically a rush to set up new shops in new places.

One high-profile problem was companies that threw together the catalogue and payments functionality in a hurry, and left fulfilment as an afterthought. One retailer pre-sold an awful lot of stuff in the Christmas rush, knowing that it would be March before some customers saw their purchases. The CEO basically said in an interview that he didn't care because he already had the money.

It's also important to remember that the internet has been around since the 1960s - it had already been organically growing, evolving and proving its usefulness.
*Technically* it's true that the quest for AI has been around for just as long, and has also produced some useful artifacts, so maybe a comparison on that level isn't out of line, but they're still not the same kind of thing.

This AI bubble is partly that gold-rush, and partly the latest silver bullet for productivity, but with multiple layers of evil on top.

It's likely that some of these newer technologies can be applied in genuinely useful ways. However, there's a lot of distance between "you *can* use it for that" and "it's the *best* tool for the job."

It's also important to remember that when referring to the software and related artifacts, "AI" is an umbrella term, not a specific technology. They're artifacts that have been discovered while pursuing the Holy Grail of Artificial Intelligence - and I use that term deliberately.

And this is where many people are unaware of history, or carefully not mentioning it. "AI" is not a new field that's only just started bearing fruit. This isn't the first harvest. The way has already been paved.

We've already had two AI Winters, one of which killed Lisp's prospects as a mainstream language, and I believe the term itself is derived from "nuclear winter" - it'll pass eventually, but a whole lot of things will die from radiation poisoning and cold along the way.
So that's what I think we're looking at for LLMs, diffusion models, and possibly even for tech like ML which has already proven its usefulness.

And that's the long version of "no, I don't think they're the same kind of thing" :)