Conversation

Notices

1. Embed this notice
   Simon Willison (simon@fedi.simonwillison.net)'s status on Sunday, 28-Dec-2025 01:46:56 JST Simon Willison

   Rob Pike got spammed by "AI Village", a poorly considered experiment in autonomous AI agents which has been sending out unsolicited emails to people (including NGOs and journalists) since April - I wrote up some notes on what happened and how it all works: https://simonwillison.net/2025/Dec/26/slop-acts-of-kindness/

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Sunday, 28-Dec-2025 01:46:55 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     @simon Simon, I tripped on this line:

     "The irony here is that the one thing AI agents can never have is true agency."

     Do you mean that they are unable to have true agency, or that they should not be allowed true agency?

     "True" agency feels like a No True Scotsman fallacy. I'm a little worried about that one. We're not sure humans have "true" agency.

   • Embed this notice
     Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸 (evan@cosocial.ca)'s status on Sunday, 28-Dec-2025 01:50:45 JST Evan Prodromou 🇨🇦🇺🇸🇬🇷🇵🇸

     in reply to

     • Rob Pike

     @simon also, I can think of a *lot* of things worse that a bot tasked only with "raise money" could do besides sending @robpike unsolicited and insincere praise email (sorry). I'm interested to see what other tools these agents had access to. Even with just email access, you can do some pretty horrendous stuff.