[sublinear]

> Everything you wrote above was said about PDAs in the 90s

No it wasn't. PDAs were seen as crappy little computers, but the applications were obvious because the bigger much more impressive computers were everywhere by then. There was no question about the value of personal computing anymore.

Everything regarding probabilistic AI is either about optimization or trading off costs. All such applications are intrinsically and perpetually lost in the weeds. The use cases aren't new because "generative" is a marketing buzzword desperately trying to cover up what is actually just "imitation".

AI makes what was already possible more accessible. It is useful, but not a revolution for the layperson or even most businesses apart from bridging knowledge gaps. It's a new way to search, but iterative at best. People are in awe of the money being exchanged, but are also in denial that it's almost entirely defense spending.

If it was just a matter of cost, scale, capability, etc. then why am I not allowed to own a flying car with my existing driver's license? Why doesn't everyone own full auto guns? Why do we serve horrible food in hospitals? Why do corporate offices thrive on work that technically never needed more than one person to accomplish even before computers were commonplace? The answers are all political.