AI still sucks ass, as far as I've seen. It's really impressive at times, but it keeps making really stupid mistakes. I think it's a fundamental issue where something that's just not actually able to think is stretched to purposes that require thinking. It's really impressive what they've done with the tech despite its limitations, but at the same time, you can't really trust AI for anything important. If it really matters, you must carefully review everything AI says - and I don't mean reviewing it against your intuition, I mean basically fact-checking.
And don't even let me get started on AI in software development, which is where I kind of have to use it in my job... Some aspects are quite useful, while others are absolutely miserable.







