Conversation

Notices

1. Embed this notice
   Justin 🌻 (onyxraven@hachyderm.io)'s status on Wednesday, 31-May-2023 02:20:53 JST Justin 🌻

   • Thomas 🔭🕹️

   @thomasfuchs the most frustrating thing is these threads, by who I thought were otherwise reputable accounts, doing things like 'asking chatgpt about itself' like parameters, configuration, etc. I guess some of this comes from the ability to 'jailbreak' the tokens that make up the 'query' stack, but there is some massive anthropomorphism being assigned that is absolutely ridiculous

   • Embed this notice
     Brad Rubenstein “:verified:” (bradrubenstein@infosec.exchange)'s status on Wednesday, 31-May-2023 02:25:24 JST Brad Rubenstein “:verified:”

     • Thomas 🔭🕹️

     half-broken, biased large statistical models...

     ... trained on stolen property and surveillance without consent.

     @thomasfuchs

   • Embed this notice
     Stu (tehstu@hachyderm.io)'s status on Wednesday, 31-May-2023 02:33:54 JST Stu

     • Thomas 🔭🕹️

     @thomasfuchs I'm so tired of being belittled by proponents of LLMs when I attempt to voice similar sentiments. Of course, it being the fad du jour means I'm dealing with it at work either way.