Conversation

Notices

1. Embed this notice
   Eniko | Kitsune Tails out now! (eniko@peoplemaking.games)'s status on Monday, 23-Sep-2024 01:20:25 JST Eniko | Kitsune Tails out now!

   Impressed by the number of people who think this seems like a plausible technological progression:

   1. Spicy autocomplete
   2. ???
   3. General AI

   • Embed this notice
     Min1123 (min1123@allthingstech.social)'s status on Monday, 23-Sep-2024 01:20:23 JST Min1123

     in reply to

     @eniko LLMs are currently a very sophisticated and costly way of making a replay attack (like TCP/UDP packet replay) against an audience that is already prepared to error correct anything that is given to it. Since we're all about grading on a curve when money is apparently sitting on the other side of the fence, this all kinda makes a morbid sort of sense to me.

   • Embed this notice
     Eniko | Kitsune Tails out now! (eniko@peoplemaking.games)'s status on Monday, 23-Sep-2024 01:20:24 JST Eniko | Kitsune Tails out now!

     in reply to

     For the record, step 2 would be "invent an entirely new technology from whole cloth which can actually support general AI" since, you know, machine learning models can't, actually

     GreenSkyOverMe (Monika) repeated this.