Conversation

Notices

1. Embed this notice
   Timnit Gebru (she/her) (timnitgebru@dair-community.social)'s status on Monday, 19-May-2025 07:21:50 JST Timnit Gebru (she/her)

   I'm thinking about spending some time to write explainers on some topics that people have asked me about. Some would include synthesizing other people's work, some would include our own work etc. Any other topics you'd like to add to the mix?

   • Rich Felker repeated this.
   • Embed this notice
     Jeremy Kahn (trochee@dair-community.social)'s status on Monday, 19-May-2025 07:22:23 JST Jeremy Kahn

     in reply to

     @timnitGebru

     I also see a need for coverage of

     (a) LLMs cannot do real knowledge work in practice, and why we can be sure it won't get better (though it may convert some jobs into "minders")

     (b) why do executives and "thought leaders" keep telling us it's revolutionary (it's to keep the bubble inflated and exert labor discipline)

     (c) how do they get away with this hype when there is not even ONE* profitable product line powered by any of the big LLM/AI companies

     *Nvidia doesn't count

   • Embed this notice
     Rich Felker (dalias@hachyderm.io)'s status on Monday, 19-May-2025 07:23:58 JST Rich Felker

     in reply to

     • Jeremy Kahn
     • jbowles

     @trochee @jbowles @timnitGebru I mean yeah, but I see the stock value evaporating as a bonus.. 🙃

   • Embed this notice
     jbowles (jbowles@infosec.exchange)'s status on Monday, 19-May-2025 07:23:59 JST jbowles

     in reply to

     • Jeremy Kahn

     @trochee @timnitGebru i have a child who was scary good at drawing from the age of 2.

     they worked and worked in any and all medium... but i've noticed the last year they've stopped doing it. Haven't talked yet about why... but i can't help but think generative AI is part of it

   • Embed this notice
     Jeremy Kahn (trochee@dair-community.social)'s status on Monday, 19-May-2025 07:23:59 JST Jeremy Kahn

     in reply to

     • jbowles

     @jbowles @timnitGebru

     This is one of many reasons I want the bubble to pop soon, even though a lot of stock value is gonna evaporate

     The longer the bubble persists the greater the long term damage caused by this creativity-drought

     I have even interviewed young programmers who tell me — *in the interview* — that they would solve problems by asking the AI to rewrite the code again until the bugs went away

     Makes me want to *weep* in frustration