I like AI to the extent that it can quickly solve well-worn, what I've taken to calling "embarrassingly solved problems", in your environment, like "make an animation subsystem for my program". A Qt timeline is not hard, but it is tedious, so the AI can do it.
And it turns out that there are some embarrassingly solved problems, like rudimentary multiplayer games, that look more impressive than they really are when you get down to it.
More challenging prompts like "change the surface generation algorithm my program uses from Marching Cubes to Flying Edges", for which there are only a handful of toy examples, VTK's implementation, and the paper, result in an avalanche of shit. Wasted hours, quickly becoming wasted days.
I feel the same way about those embarrassingly solved problems! Though oftentimes the trick is knowing what to ask for. I remember grinding for weeks on a front end but until I realized what the problem was (not the exact bug just what the general concept should be) Claude then fixed it in 10 seconds.
And it turns out that there are some embarrassingly solved problems, like rudimentary multiplayer games, that look more impressive than they really are when you get down to it.
More challenging prompts like "change the surface generation algorithm my program uses from Marching Cubes to Flying Edges", for which there are only a handful of toy examples, VTK's implementation, and the paper, result in an avalanche of shit. Wasted hours, quickly becoming wasted days.