I'm not convinced they are throwing money at nothing this time. If you look into my post history (don't, I'll give you a tl/dr) you can see me being fairly critical about the state of LLMs, but also that I find them incredibly useful for some things. GPT basically writes 95% of our documentation now, only the code we absolutely don't want to share with the internet is still be documented by hand. And that is just a tiny bit of how the new AI boom is actually useful. My wife works with dyslexic people, and the amount of usability AI has in terms of turning the written word into audio is outright amazing. I'm sure this will be hard on a lot of voice actors, copy writers and so on, but AI can do those jobs and a lot of other things.
To me that means that, unlike blockchain, there is actually a viable product behind all the AI hype, and unlike Uber companies have already started selling it rather than waiting for it to become profitable through scale.
Maybe it's just me but the AI hype reminds me a lot more of the dotcom era than any of the recent hype cycles. Of course this means that for every Amazon or Google there will be 9 million failures, and unlike the dotcom era, it's likely that it'll be Google (or most likely Microsoft) who hits the jackpot, meaning it'll be even harder to find the right investment opportunities this time, but I don't think it's a fad.