Conversation

Notices

1. Embed this notice
   scribe (scribe@mastodon.sdf.org)'s status on Friday, 13-Sep-2024 18:16:40 JST scribe

   "The machine also dealt with its human interlocutors in polite and empathetic terms, in contrast to the scorn sometimes heaped on conspiracy theorists in real life."

   Fascinating (but pay-walled) article on how chat #AI can influence people away from conspiracy beliefs. Humans are messy and subjective. Is an opaque summary machine more reliable and more engaging when it comes to discerning truths?

   https://www.ft.com/content/909f33d1-0d33-440b-a935-cfa65d2fccd1?segmentId=b0d7e653-3467-12ab-c0f0-77e4424cdb4c

   • Embed this notice
     scribe (scribe@mastodon.sdf.org)'s status on Friday, 13-Sep-2024 18:18:37 JST scribe

     in reply to

     Follow-on thought: How close is our use of #AI to gaming?

     Is there an intersection around being able to *play* and *experiment* with a machine that is hard to get with other humans? Are chatbots seen as a kind of "safe space"?

   • Embed this notice
     scribe (scribe@mastodon.sdf.org)'s status on Friday, 13-Sep-2024 18:41:39 JST scribe

     in reply to

     • \\\\\\\\\\\

     @11backslashes Effectively, yeah. I think the big swing is that modern AI feels like it's answering your questions, whereas Eliza was all about being the questioner to open up your own thoughts.

     Still, AI therapy I think is probably an untapped market, if it can do some kind of basic empathy. (which may be different to being an *appropriate* form of therapy... ;)

   • Embed this notice
     \\\\\\\\\\\ (11backslashes@social.unx.pw)'s status on Friday, 13-Sep-2024 18:41:40 JST \\\\\\\\\\\

     in reply to
     @scribe It's ELIZA all over again, but with significantly better modelling.
   • Embed this notice
     scribe (scribe@mastodon.sdf.org)'s status on Friday, 13-Sep-2024 18:42:17 JST scribe

     in reply to

     • Fionnáin

     @ephemeral Yeah, it could be cool to have an option like "click here to discuss these topics further" in an interactive way!

   • Embed this notice
     Fionnáin (ephemeral@mograph.social)'s status on Friday, 13-Sep-2024 18:42:19 JST Fionnáin

     in reply to

     @scribe interesting idea. Like an automatic tl;dr that just says 'article makes unreliable claim that aliens have replaced humans' or something. It would be nice if there was some potential in this.

   • Embed this notice
     scribe (scribe@mastodon.sdf.org)'s status on Saturday, 14-Sep-2024 00:02:59 JST scribe

     in reply to

     • Justin Pickard
     • D-ro T. M. Sullivan 🍜🍵

     @jcalpickard @tmsullivan Chaos theory prompting?

     Sounds good though. How to ask a question of the I Ching has always come down to a disruptive process for me - if it's a disposability in, disposability out. Maybe we need more mindful prompting UIs.

   • Embed this notice
     Justin Pickard (jcalpickard@assemblag.es)'s status on Saturday, 14-Sep-2024 00:03:00 JST Justin Pickard

     in reply to

     • D-ro T. M. Sullivan 🍜🍵

     @scribe Working on a thing with @tmsullivan about "ostensive detachment" in LLMs and divination; I think there's also something about the iterative input-output tinkering, and being able to (quickly) see the effects of minor tweaks and changes.