"Normal tech" puts people out of careers every so often, and kinda randomly. Society has benefited as everyone who isn't put out of a job significantly outnumbers those who are.
I was expecting AI to put me, a software developer, out of a job after everyone else, because I thought coders would be needed to write the AI for everyone else, one job task at a time. Right now, it's looking like LLMs can translate not only English into Spanish, but business requirements into code. Sure, not perfectly, but well enough that I feel like I'm living that joke about a mechanic who charges $1 for a new bolt and $999 for knowing which bolt. #LeopardsEatingPeoplesFacesParty2026, but seriously, LLMs have been able to train new AI for a while now.
I hear similar from professional clothing photographers who find that even if they want to support the human models, they're forced by economics to use the GenAI models. I hear rumblings of problems from copyeditors, translators, and mathematicians. It's affecting a lot of people at the same time, even if they can all point to ways it kinda still sucks or "lacks taste" or "has no common sense" or whatever. I can point to the models doing bad code, but I know from experience that "bad code" can win awards at the product level.
Robots have been milking the cows and sorting the post for decades now. But even the first agricultural revolution had a lot of people moving out of the country and to the cities because they could no longer work on the land.
Ironic that I only hear praises of AI-driven cars from one place, which has been saying "just a few years" for the last 13. The economic impact if that actually ships without any caveats would also be huge (which is something I'd want people to keep in mind if they feel like responding with how good Tesla's current version is, I've heard that for the last three versions at least).