pull down to refresh

IME they struggle with the compression step a lot more than they do generation.

It took me 1 hour of prompting to generate the code I've spent 16 hours compressing.

Because you can't really "compress" code much, right? You need to actually re-architect and provide standards relevant to your project.

Or, you need to include more design requirements in your initial plan. If you work too organically, you end up with a mutant creation.

Me, I have a team to make things real. So my vibes live as R&D prototypes, not production code. This way we can encounter all the problems and decisions once before doing it "for real."

reply

Compression and rearchitecting are synonymous in this case - architecting to reduce surface area for human consumption is relatively foreign to bots.

reply

another thought is whether compression is also harder than generation for humans as well

I think it probably is

reply
150 sats \ 0 replies \ @k00b OP 5 Apr

It definitely is. I call that out in the post. But bots don't need compression to make progress as much as humans do.

reply

I wonder how much their performance depends on the need to reason about state vs reasoning about the code itself?

reply

Statelessness is easier for humans and bots alike, but I do think the generation bias is legit independent of the context.

Anytime I’ve prompted “make this clearer” or “clean this up,” they tend to increase lines of code. Even “reduce lines of code” results in, at best, negligible reductions. I have to point out excessive abstraction and overengineering repeatedly.

reply