> When a bubble happens, it's usually preceded by the companies involved reaching obviously inflated evaluations, which makes the burst obvious in hindsight.

I think this highlights a big issue with both sides of the "AI is/isn't a bubble" debate - the idea that there's an established pattern that economic bubbles cohere to. There really isn't, and economists haven't been able to find a consistent pattern that past bubbles have conformed too (that isn't also the case of non-bubbles). "Bubbles for Farma" is a cool paper that tries to do this[0].

Like everyone else, I have my own opinions on whether there is/isn't a current AI bubble. But we should all stay humble and try and avoid stating anything as if it's an absolute fact.

[0] https://www.nber.org/system/files/working_papers/w23191/w231...