Post #1966240
2026-02-08 13:03 UTC
@xjki reviewing AI code is as exhausting as it is absolutely necessary. One reason it’s exhausting is that I can’t use any of the mental shortcuts I use when reviewing hand-written code. I can form a mental picture over time of the developer and what they’re thinking, their style, their strengths, what they misunderstand. I can focus on areas they are likely to have made mistakes, and skim the areas that are probably fine. But AI isn’t like that, and you have to be on guard at every turn.
Replies (3)
-
@cocoaphony@mastodon.social 2026-02-08 13:07
@xjki and it’s not that the AI code is universally bad. It’s actually usually pretty good. But it goes off th rails in such unexpected ways, and sometimes subtly where the code *looks* good but has actually completely broken something, sometimes for understandable reasons, and sometimes bizarrely. If it were a single developer, you’d never believe they both wrote code section A which is excellent and meticulous, and code section B which is sloppy and makes beginner mistakes.
-
@xjki@mastodon.social 2026-02-08 13:15
@cocoaphony It absolutely is. It's also existential dread. Am I training robot to be even better (yes, I am most probably)?
-
@jasonhowlin@mastodon.social 2026-02-08 15:25
Agreed. I feel like I’m always cleaning up after it in code reviews.