Cognitive Burden
(kau.sh)4 points by codeclimber 10 hours ago
4 points by codeclimber 10 hours ago
i sense a slight strawman argument here.
you pick a subset of admittedly clear problems with how aigen is being leveraged but blow it up to be the general case for everyone.
you're quick to call into question the motivation of programmers who use aigen.
i assure you when you see people like Mitchell Hashimoto using aigen, you owe it to yourself to at least explore the possibility that there are folks building great software, and product using aigen. I'm not saying you need to be convicted but your comment comes out as being one sided and emotional.
i'm also curios how you square your thoughts around actually building good products. there's plenty of bad apples on all sides (like the craftsmen who tdd apps that land up not being used by anyone).is that true value?
it would be silly and a strawman argument to then assume that to always be the case.
The author claims lower cognitive burden from debugging ai output than from doing the work themselves - to which I'd counter that they're probably in the wrong profession if the process of the craft itself is tedious to them. There are plenty of writers who love the act of writing, and I'm sure they'd feel like I do about aigen code tools - it's way more effort, for me at least, to debug generated code with subtle (and sometimes not-so-subtle) bugs than to iterate it out myself. If I don't know what direction to take, that's what TDD gives me - build the next smallest part, forget about the big picture for a while.
And if this is the sentiment of someone who is not impassioned with their craft, then I suddenly understand the ai uptake for codegen too - so many people are in programming professions for the simple reason that it can pay quite well, not because they actually enjoy solving problems with code. Now, I respect your right to select for money over happiness, but this has brought inherent problems in the past - coders who don't care are sloppy, and I've had to clean up after a few of them. Coders who don't care _and_ use aigen are more likely, therefore, to produce reams of code they have no interest in, no desire to see crafted "just so" - it's a job, a 9-too-5, a paycheck. Which is fine for them, but increases the burden on their team as the people who _do_ care have to pick up the pieces.
I also see a lot of what can only be described as an entitlement - expecting to produce production-ready code outputs without knowledge of how any of it works. This is an even sharper protrusion of the scenario outlined above - these are people bold enough to openly state that they don't care about the craft or the process, and only care about "productivity", where "productivity" is measured in commit and pull request counts, conveniently leaving out defect counts or user impact.
I can't wait for this ai bubble to blow wide open.