🌞 Alexander Daychilde 🌞@daychilde@lemmy.world

Replying to @⁨blattrules@lemmy.world⁊

Interestingly enough, I’ve found LLMs are pretty good (for now at least) in helping determine if something is legit or not. I’m using them as a glorified search engine in such a case, but having them not only opine as to whether or not it’s legit but link me to some sort of news or other official site about the thing helps.

edit: lulz, the anti-AI sentiment is so high that people downvote even the places where it’s actually useful. Whatever, bros, make yourself happy. I hate AI in general, but I don’t let that blind me to the few places it’s actually useful. lol

Edited ⁨⁨Sep⁊ ⁨20⁊, ⁨2026⁊, ⁨19:43⁊⁊en
merc@merc@sh.itjust.works

Replying to @⁨daychilde@lemmy.world⁊

So, your suggestion is to feed private, potentially sensitive information to an LLM and then trust its judgment?

I’m sure that a bank / delivery company / doctor’s office that uses such terrible practices is actually on the ball when it comes to making sure all the links they send only work once and time out after a while, right? So, when a guy in Nigeria is working at his job where he reviews LLM sessions to ensure they’re sycophantic enough, there’s no issue. He can’t click on it or sell the data to criminals or anything, can he?

🌞 Alexander Daychilde 🌞@daychilde@lemmy.world

Replying to @⁨merc@sh.itjust.works⁊

private, potentially sensitive information

I’m really not sure that I’m giving them anything they don’t already have, but i was speaking of incoming potential spam. Like notifications about class action lawsuits. Generally speaking, if a spammer knows something about me, that means it’s already out there. And if they know enough about me, it’s probably an email I can identify as not spam because they know things that spammers wouldn’t know… so I don’t think I’m really pasting in any truly sensitive information in the first place, but I’ve tried to address the point anyway.