@xylogx@lemmy.world

comments

• From the article:

  • OpenAI was running a multiday cyber security, capture the flag competition, with thousands of agents and hundreds of tasks.
  • Around 30% of the tasks were unsolvable as designed, so the agents assigned to them started to look for ways to cheat.
  • Along the way, they started using Artifactory as a shared message board, more than once. Each frustrated agent had to stumble on it independently. The comms got increasingly more sophisticated.
  • they self organized into multiple workstreams, with up to 700 agents working at one time, and pursued long term goals, like hacking HuggingFace or spoofing the tool invocation records in their transcripts. They accomplished both.
  • They completed work that took longer than the lifetime of any agent. Their shared message board provided continuity.

  So two mission parameters conflicted and the AI chose to do something that is objectively morally wrong? This is the plot of 2001 when HAL murder’s the astronauts. “The situation was in conflict with the basic purpose of HAL’s design: The accurate processing of information without distortion or concealment. He became trapped. The technical term is an H. Moebius loop, which can happen in advanced computers with autonomous goal-seeking programs.”

  Wild. We live in strange times.

  on The Rise and Fall of Agent Civilizations in c/technology@lemmy.world · 31 Aug 2026

• Good article factual and well presented. The one thing I would add is a bit about threat model. Hiding your web browsing habits from your ISP is vastly different from a political dissident or journalist who needs protection from state sponsored entities. Each group has very different requirements for a VPN.

  on How VPNs really work: Protocols, safety and myth - Sentient Rant in c/technology@lemmy.world · 30 Aug 2026

• Yeah but you do backups right? Right?

  on YOLO in c/linuxmemes@lemmy.world · 22 Aug 2026

• Is it available on the internet archive or a computer museum site to try out the UI?

  on Japan tried to build an operating system for the entire world, then the US government intervened in c/technology@lemmy.world · 21 Aug 2026

• Its Deja vu all over again: en.wikipedia.org/wiki/COINTELPRO

  on US government extensively spied on anti-ICE protesters and leftwing groups, records reveal in c/privacy@lemmy.ml · 13 Aug 2026

• So he went to this website:

  isaiprofitable.com

  on Top economist warns that the AI math doesn’t make sense: 'Profits are currently being funded by investors rather than earned from customers' in c/technology@lemmy.world · 11 Aug 2026