Conversation

Notices

1. Embed this notice
   akkoma flavored lizzy (fleckenstein@lizzy.rs alt) (lizzy@social.vlhl.dev)'s status on Friday, 21-Mar-2025 10:02:27 JST akkoma flavored lizzy (fleckenstein@lizzy.rs alt)

   people are so brain damaged by LLMs

   i just watched a german streamer who wanted to know the distribution of seats in the german parliament after the election ask chatgpt about it, notice that they gave the FDP 10% of the seats (they have 0), then ask ChatGPT for a correction, it admits the mistake and then gives the exact same results again, then he tries to do mental math to account for the obvious mistake while trusting the rest of the data

   bro literally just look it up on a search engine or wahlrecht.de what is wrong with you

   • Johnny Peligro likes this.
   • Embed this notice
     Johnny Peligro (mischievoustomato@tsundere.love)'s status on Friday, 21-Mar-2025 10:02:48 JST Johnny Peligro

     in reply to
     @lizzy
     > we should ban chatgpt
     i don't like this kind of "solution" at all, teaches people to rely on the government. I do not like that
   • Embed this notice
     akkoma flavored lizzy (fleckenstein@lizzy.rs alt) (lizzy@social.vlhl.dev)'s status on Friday, 21-Mar-2025 10:02:49 JST akkoma flavored lizzy (fleckenstein@lizzy.rs alt)

     in reply to
     we should ban chatgpt like what the fuck people can't even use a search engine anymore, much less remember domain names or use their fucking brain
   • Embed this notice
     D:\side\ (dside@mastodon.ml)'s status on Friday, 21-Mar-2025 10:03:13 JST D:\side\

     in reply to

     @lizzy I think at some point the sturggle becomes entertaining enough to press on. And it can be difficult to distinguish whether they're genuinely looking for an answer still or at that point just messing with the machine just to see what will happen. With streamers being in large part entertainers the line is probably very blurred.

     I was pointed at a similar example recently where ChatGPT very confidently lied about its own generated images and its supposed inner workings, and the host just… believed it? Despite confidently catching it lie minutes before? And the worst part is I don't think it was for entertainment, they seem to be drawing or at least delivering conclusions with a straight face.

     https://youtu.be/160F8F8mXlo

     Johnny Peligro likes this.
   • Embed this notice
     Ian Molton 🇺🇦🇵🇸🏳️‍🌈🏳️‍⚧️ (spyro@fosstodon.org)'s status on Friday, 21-Mar-2025 10:03:50 JST Ian Molton 🇺🇦🇵🇸🏳️‍🌈🏳️‍⚧️

     in reply to

     @lizzy it has it's uses. It's great for a quick first order guess, but it's so utterly surrounded by problems that i agree. As implemented, it ought to be banned.

     They are lying to people about his well it works, and people are so utterly uncritical in their thinking. A recipe for exactly the kind of disaster 'they' want.

   • Embed this notice
     Johnny Peligro (mischievoustomato@tsundere.love)'s status on Friday, 21-Mar-2025 10:03:50 JST Johnny Peligro

     in reply to

     • Ian Molton 🇺🇦🇵🇸🏳️‍🌈🏳️‍⚧️
     @spyro @lizzy for me, AI has been great to solve issues for me or at least direct me/make me think in a more guided way to fix my issues on my own. It's definitely been helpful.