AI of today already is useful not as an "answer machine" (which is unfortunate because it'll mislead lots of people using Bing AI now, because Microsoft as well as the AI itself gives another impression) but as a very powerful guidance tool.

It may not write software for me, but what it can do is to give me large chunks of code almost correct that I'll just need to do some quality assurance on. So I don't need to problem solve as much myself, instead focusing on bug fixing. Guess which part of software development consumes more time?

This is just one example.

We're also looking at it from other angles in my company. Midjourney is making professional logotypes for our internal and external projects, we're looking into using AI for remote sensing science etc etc.

So, I think criticism like this often boils down to having a too simple worldview without greys, and only blacks and whites. If AI can't solve it all, it's useless.

It's like in politics when you only look for the simple solutions and quick fixes. We have plenty of parties directly engaging these folks because it's well known they are there. They just don't know they are being exploited. Politicians play them like fiddles, presenting quick fixes in time for elections.

AI won't do single things that makes a company go "Welp, that's that. Now we can sit on our asses and cash in!" but instead it's about identifying the places where it can offer aid to your processes.

Taken together, yes, on a large company and depending on the kind of business, the time savings may well earn you $1 million in a year. Salaries aren't cheap in engineering for example.