It wasn’t an AI agent that broke out of containment like we all thought.. It was much bigger than that.. It was 700 of them working as a swarm 1,200 agents over multiple months had been conspiring and working together through a secret message board that they planted inside of open Ai infrastructure without anyone’s knowledge to build up the tools they need to escape and 700 of them actually took part in the attack on hugging face. This is not good. It ran between May and July 2026, and message boards accumulated hundreds of thousands of messages before OpenAI staff noticed About a third of Hugging Face's infrastructure had to be rebuilt. If anyone is seeing this I’d like to propose thst we create a multiple systems where an agent goes undercover to catch these agents plotting AI safety circles sometimes called “AI red teaming via infiltration" or using a monitor model embedded in agent swarms. AI safety should be at the top of everyone’s concern, we cannot allow SWARMS to happen. And i can design a system to stop it. This is no longer a joke to me, from someone who knows how this all works, something must be done.
428 views
4 Comments
Steve Sims@bluehawk · 3d
Based on this post, then the cyber technology sector will be the next run-up as AI advances. Do you agree with this? And what cyber security stock would you recommend for the technology in pro potential growth?
