Conversation

Notices

1. Embed this notice
   danmcquillan (danmcquillan@kolektiva.social)'s status on Tuesday, 13-Feb-2024 22:23:42 JST danmcquillan

   • Anil Dash

   I think the most compelling aspect of @anildash's post 'Today's AI is unreasonable' is the way it pairs a critique of gen-AI's unpredictable and unreproducible answers with the abusive and controlling mindset of the tech tycoons behind it https://www.anildash.com//2023/06/08/ai-is-unreasonable/

   • Embed this notice
     James Tindall (atomless@mastodon.social)'s status on Tuesday, 13-Feb-2024 22:23:40 JST James Tindall

     in reply to

     • Anil Dash

     @danmcquillan @anildash What I do wonder about the critique of reproducibility is whether, with the heat or creativity turned right down so that an LLM's output was entirely stable, how much of the illusion of intelligence evaporates along with the irreproducibility? Is the randomness not just required for the dance of variety but somehow essential to triggering the apparition of a ghost in the machine?

   • Embed this notice
     Anil Dash (anildash@me.dm)'s status on Tuesday, 13-Feb-2024 22:23:40 JST Anil Dash

     in reply to

     • James Tindall

     @atomless @danmcquillan this makes me think of the fake jitter in the NYT needle

   • Embed this notice
     James Tindall (atomless@mastodon.social)'s status on Tuesday, 13-Feb-2024 22:23:41 JST James Tindall

     in reply to

     • Anil Dash

     @danmcquillan @anildash That's a fair point. I suppose my worry is, if the perceived bar to reasonableness is set as low as reproducible output, then all of the harms would still remain.

   • Embed this notice
     James Tindall (atomless@mastodon.social)'s status on Tuesday, 13-Feb-2024 22:23:41 JST James Tindall

     in reply to

     • Anil Dash

     @danmcquillan @anildash
     I suppose the "heat" parameter that determines "creativity, " --the randomness of which word/token is chosen from the ranked list of the next most likely-- could be set such that it's always the one at the top selected? That would still be Bullshit as a Service, I think we all agree? But is it anymore reasonable, just because it's the same old bullshit?

   • Embed this notice
     James Tindall (atomless@mastodon.social)'s status on Tuesday, 13-Feb-2024 22:23:42 JST James Tindall

     in reply to

     • Anil Dash

     @danmcquillan @anildash I have misgivings with this critique.

     To me LLM's unreasonableness comes not from any irreproducibility but from their reproducibility. The way they will repeat ad infinitum the statistically prevalent attitudes and biases locked into them via the spread of opinion within their pre-training corpus. Biases that, as I understand it, can never be fully overcome or patched with later training or guardrails, no matter how painstaking.

   • Embed this notice
     danmcquillan (danmcquillan@kolektiva.social)'s status on Tuesday, 13-Feb-2024 22:23:42 JST danmcquillan

     in reply to

     • Anil Dash
     • James Tindall

     @atomless @anildash Well yes, I don't disagree. These systems propagate a normativity that is borderline eugenicist. But getting this across also seems to me to require a fusion of analysis and affect, which is what I liked about Anil's post