Bubble implies exuberance and overvaluation of financial assets, not whether the technology is valid or not.

It's easy to see how many current AI companies could end up being dud investments, even if they produce good models

> AI isn't a bubble [because] AI will deliver just like the internet.

And everyone knows there was no internet bubble. No period of proliferation of available venture capital, rapid growth of valuations followed by all gains being lost.

> Bubbles may form when enthusiasm and money gets ahead of progress

Nah, it will deliver on some of those promises but there is definitely zero chance that it will deliver on all of them (maybe even most of them), but this also isn't how bubbles form.

I think the Internet is the most apt example.

The Internet delivered 5 trillion dollar companies. It was also a bubble that popped in 2002. For every Google or Amazon there are a dozen Nortel's and Webvan's.

internet was both a monumentally successful technology, and a bubble. The dot-com crash of 2000 saw most stocks give up all gains from the prior ~5 years, the surviving stocks took over a decade to recover their previous high.

AI bubble will deliver just like the internet bubble.

Eh, I'd change the analogy.

AI is a bubble in the sense there isn't a direct path of put more money into research and get a better result. I'd almost argue we are spending money in the wrong areas.

The internet had clear and well defined steps for improvement. Although we can increase accuracy or whatever by 1-2% that doesn't define definitive steps toward AGI.

Humanity is spending money in a lot of different areas, and we're seeing both theoretical and functional progress. The progress in applications is already destabilizing a lot of industries in just the last year, how much faster should we progress?