True, but it is impossible to catch up while preserving quality and mental sanity.
I know about several of my friends, non-tech, being directly impacted by AI.
In finance, lots of analysis work is now offset to LLMs, and the people leveraging the tools obviously still have the issue that they need to review everything the AI has analyzed, their formulas, etc. And lots of nuance and things that a human would caught are lost. But in the meantime the expectation is that your analysis output is 5 times what it was before.
My girlfriend works in corporate law for an insurance company. The company is FOMOing hard for LLMs and pushing everybody to write gemini "gems" and notebooklm presets to do lots of the work.
But it absolutely does not scale: you can't keep up with those demands, while also providing the same quality coming from thoroughly analyzing new regulations and such.
Another friend that works in credit has now the company mandate that people update financial statements etc directly to LLMs and those tools come with a yes/no about whether they will finance it or not. Quality of debt has now plummeted, needless to say and the process is longer that it has ever been because re-reviewing the LLM analysis is more expensive than doing it on your own.
My own bank has had a terrific customer care that has been recently replaced by an LLM, tragedy. It is absolutely unhelpful beyond the 80% pareto principle where customer care had already pre-canned answers anyway. But for the 20% of cases that are major issues/bugs, the AI is simply not helpful.
My bank genuinely had a bug with invoice processing and there was no way to tell them nor to resolve my issue (which required somebody to manually void the previous invoice and restart the process that got bugged).
I think it's a tragedy.