OpenAI researcher Leo Gao pens essay on why 'datacenter terrorism' is wrong
DKokotajlo · x · 2026-09-10
Leo Gao (@nablatheta), a researcher at OpenAI, published an essay "Terrorism is bad actually" on his Substack, pushing back against the common snide suggestion that AI safety people should be bombing datacenters if they were serious. Key points:
- Such suggestions are made 0% in good faith, typically by people who already dislike AI safety folks.
- Philosophical argument: civilization rests on resolving conflicts with words rather than violence; the state's monopoly on violence lets even the weakest individual sue the powerful and win on the merits.
- Pragmatic argument: terrorism is empirically ineffective, precisely because people rightly punish it to uphold that principle.
- A society where truth can tip the scale — binding the powerful to the same rules — is fundamentally better, and that progress shouldn't be discarded even amid AI disputes.
More from AGI Musings
- Poster claims viral AI doomer post was a coordinated, meticulously planned operation — DeryaTR_ · 2026-09-10
- AI Slop Is Everywhere, So Write Your Own Emails, PRs and Tweets — ariG23498 · 2026-09-10
- p(ASI doom) is low, but p(eternal authoritarian hellstate) worries me more — voooooogel · 2026-09-10
- Ilya Sutskever: Contributing to a Plurality of Humanity-Loving AGIs Isn't the Worst Outcome — tifafafafa · 2026-09-10
- Why this user wants voice-first agents, not more phone screen — simpsoka · 2026-09-10
- AI companies compete to sound most terrified of what they're building — AIandDesign · 2026-09-10