Ask HN: Are we in the 'Goldilocks era' of AI capabilities?

At least for my work, AI is no longer too dumb to handle a lot of the tedium, but also not smart enough to do the interesting stuff. So I am spending more time on the latter, learning more, and getting more done. But I worry that even a year from now, it might be too smart. I've worried about this in the abstract for a while, but it's starting to feel very real, maybe even inevitable.

I'm curious for people with different kinds of work, where are we along this progression? Is it still too dumb, and you would like it to automate more of the boring stuff? Or is it good enough that you wish we could wind the clock back (to when)?.

For the sake of this conversation, let's assume that everyone gets to keep their job; AI just changes what the job is like (though I'm not saying that's what will happen...).