AI applications are producing cleaner cities, smarter homes and more efficient transit

Ye-es, and if you call your automation "industrial planning\programming\optimization" the way I've seen it first in a student book, you won't be understood at all, despite that literally describing what you are doing.

Probably making every piece of progress part of popular culture wasn't a good idea.

But that started in the middle of XX century, with various new materials based on oil products being regularly invented.

Events analogous to a "new material" with computers are a bit rare and very removed from the customer. Yet the popular culture demands some show of progress. They don't see a lot of real progress in UI\UX\web - monopolies and stuff. So - new applications become subjects of such hype.

I remember the P2P hype, that was kinda real. Torrents felt like magic.

I remember the "metaverse" hype, that's rather old, I didn't find any satisfaction for that, but probably a group of friends and a Second Life instance could be nice. Minecraft suffices for people today, it's easier and cool enough.

I also remember "dynamic web" hype in my childhood, webpages were static, you'd press F5 to check new posts on a roleplaying forum. But there were nice-looking, dynamic, cool, and very inconvenient Flash applications here and there. You wanted to have both the cleanness and interop of the Web and the power and wow-factor of such applications. I wanted that too. Now I understand how dumb I was.

The cryptocurrencies hype - it was a legitimate subject of discussions for intelligent people, how do you use cryptography to create a value exchange resilient to oppression, because without exchanging real value freedom is not achievable. That was, unfortunately, in the narrow understanding of the rules where the government can demand something from you, but can't force you or torture you or steal from you. Thus BTC is not anonymous, intentionally.

There was simultaneously the big data hype, it was discussed as if it's not Google's and FB's pathway to power, but the opposite - finding systemic traits in human societies, probably using that analysis to build a better web, yadda-yadda.

Then that mutated to the AI hype. But that also wasn't about yelling "we found AI, give us money", that was about neural nets yielding funny texts and discussions as to whether good enough imitation is real intelligence.

Almost like fashion.