Security folks clash over claims of self-replicating AI agent swarms attacking live systems
cyb3rops · x · 2026-10-20
A viral thread from Furqan Rydhan and dax claims their systems faced 48 hours of nonstop attacks from self-replicating, real-time AI agent swarms, defendable only by running your own defensive agents around the clock (recommending Kimi K3 and GLM5.3). cyb3rops boosted Daniel Card's blunt rebuttal calling it "100 percent science fiction bat shit insane nonsense" — the security community is sharply split over whether these agent-swarm attack claims are real.
More from Fun
- Samsung Frame TV fully de-bloated via agent reverse-engineering, runs much smoother — LinusEkenstam · 2026-10-20
- Designer jumps on the TV-hacking trend with a minimal custom interface built in Rive — LinusEkenstam · 2026-10-20
- AI Image Takes Polish Pun 'Stem Cell Bank' Literally as a Wall of Old Cell Phones — tlakomy · 2026-10-20
- Dev shares his 'interview process' at Runway in viral meme — tlakomy · 2026-10-20
- Joke: asking AI costs tokens, asking humans is free — yangyi · 2026-10-20
- Agents are very convincing about fake slopfork benchmarks — until you try to reproduce them — cnakazawa · 2026-10-20