The curl project has been bombarded by large volumes of low quality AI slop security reports and Daniel shows examples. Sloppy humans causing Denial-of-Service attacks by overloading maintainers with quickly produced almost-real-looking rubbish.

At the same time, upcoming new AI powered tools find flaws and mistakes in existing code in ways no previous code analyzers have been able to. Daniel names names and shows examples of findings, some that even feels almost human. Next level bug-hunting for sure.

AI now simultaneously brings us the worst and the best.