News for hackers. Stuff that matters.
Re:Hacker plants false memories in ChatGPT to steal user dat... by rdtsc on Wednesday September 25, 2024 @08:34AM (#41646683)
Seen this with various users jumping into GitHub issues, replying with what seem like well written, confident, authoritative answers. Only looking closer, it’s referencing completely made up API endpoints and settings.

It’s like garbage wrapped in a nice shiny paper, with ribbons and glitter. Looks great, until you look inside.

It’s at point where if I hear LLMs or ChatGPT I immediately associate it with garbage.

Re:Hacker plants false memories in ChatGPT to steal user dat... by MichaelZuo on Wednesday September 25, 2024 @06:50PM (#41652687)
However, it is a handy way to tell which users have no qualms about being deceptive and/or who don't care about double checking.