Yeah it feels like it performs really well if you know what needs to be done and how. Spare no detail from the description of how exactly it should do it and it'll usually get most of it right and save some of the time spent messing around with initial setup and boilerplate.

That's were I've had my best luck with it.

You could say buggy output is one of the main problems with neural nets. In fact all bugs on earth exist because of neural nets.

The most mind blowing thing to me is how it’s responsive to feedback like this. It’s like pair programming with a junior. It might be faster for me to just type it myself, but it’s doing what I tell it and filling in a fair amount of blanks itself.

But is it responsive? If I gave ChatGPT the same 6502 prompt now that it's done it once would I also have to correct it about the addition as well?

You are not asking for something impossible, its just not there yet. The speed of progress in the field is something few can extrapolate well over next 5 years.

It “remembers” the context of prior prompts so the answer is, in the same chat, “sometimes.” Other it doesn’t seem to bring enough weighting.