Note: 29

Reading this HackerOne report really made me feel the harm being caused by the endless stream of AI slop.

This marks the first time (that I know of…) where I’ve come across a bug report online and didn’t immediately flag it as AI. I was a bit confused while reading the code, but was just skimming it, and it all seemed plausible. But then I get to the comments, and people are saying that the repro code doesn’t cause a crash as expected, that the code in question actually works fine, and I realize: damn, I got got.

I don’t know how long it took @bagder@mastodon.social and team to test this, and I don’t know how many of these they get per day (at least 18 reported so far: https://gist.github.com/bagder/07f7581f6e3d78ef37dfbfc81fd1d1cd), but, god, what a waste of time.

Any input form with a potential reward is being spammed with convincing sounding nonsense, from bug bounties, to every student’s homework assignments, to the Clarkesworld Magazine: https://neil-clarke.com/how-ai-submissions-have-changed-our-submissions-process/.

The root problem is wealth inequality of course: everybody needs money, most people don’t have enough of it, and AI slop is a low effort way to get it.

Until the day we solve capitalism though, we gotta wade through the muck.

Keep up the good work, badger et al, I hope you’re figuring out ways to keep your sanity intact!