Mollick: the Mythos security freakout and limited cyber-model rollout were effective risk mitigation

emollick · x · 2026-11-17

Ethan Mollick argues that the much-derided Mythos security freakout, along with labs' practice of first deploying cyber-capable models with limited access, has effectively mitigated AI's downside cyber risk.

He adds that every outage over the past five days has been seized upon by people who previously ignored cybersecurity, each implying "this is the big one," fueling panic and rumor-spreading on X and in real life.

Related event: Mollick: Fragile Infrastructure Crashes on Its Own, No AI Attack Needed(4 posts)→

Original post →

More from AGI Musings

AGI Musings channel →