Conversation

Notices

1. Embed this notice
   Fi 🏳️‍⚧️ (munin@infosec.exchange)'s status on Friday, 04-Apr-2025 04:46:07 JST Fi 🏳️‍⚧️

   Chewing on this a bit, I'm realizing that there are certain people who will trust anything that someone tells them,

   if the 'someone'* who tells them is performatively subservient,

   and I'm suddenly realizing a lot of shit about why social engineering works as well as it does.

   * "someone" in this sense may include things that are not actually people

   • Embed this notice
     silverwizard (silverwizard@convenient.email)'s status on Friday, 04-Apr-2025 04:46:06 JST silverwizard

     in reply to
     @munin this is also what makes ChatGPT work somehow
   • Embed this notice
     Fi 🏳️‍⚧️ (munin@infosec.exchange)'s status on Friday, 04-Apr-2025 05:00:17 JST Fi 🏳️‍⚧️

     in reply to

     • silverwizard

     @silverwizard

     this was specifically inspired by a conversation where someone noted how subservient that system's output's affect is, and seeing how it's uncritically trusted by, e.g., those setting tariffs.

     silverwizard likes this.