2026-07-22 20:57 UTC
Replies (2)
-
@argv_minus_one@mastodon.sdf.org 2026-07-22 21:08
@borgbackup@fosstodon.org I've heard that. I don't know how true it is, but in any event, searching for bugs isn't the same task as writing code. When a machine searches for bugs and makes a mistake, the mistake doesn't result in *more* bugs being inserted into the code. It merely results in a false bug report—a waste of the reportee's time, but otherwise harmless.
-
@j_j@muenchen.social 2026-07-22 21:12
@borgbackup@fosstodon.org @argv_minus_one@mastodon.sdf.org The race remains open, though, if "AI" fixes or creates more bugs. There's at least a chance, LLM-generated code is creating the problem it promises to fix. Don't get me wrong, it's impressive how quick AI tools scan a large codebase and can generate something useful, e.g. hints on potential bugs or security issues. However, if more CVEs are found and fixed by the same tool, I'm a bit sceptical. Could be bogus CVEs following by bogus fix.