Belfer paper adds 8 more predictions across closed- and open-weight harm pathways

StephenLCasper · x · 2026-10-20

Thread follow-up: besides the central prediction, the Belfer paper makes 8 further falsifiable predictions — three on the closed-weight harm pathway, three on the open-weight pathway, and two applying to either.

Related event: Harvard Belfer Paper Warns First Major AI-Enabled Terror Attack Could Come as Early as 2027-2028(10 posts)→

Original post →

More from Safety

• Joshua Saxe: Curb AI cyber misuse without blocking open weights — joshua_saxe · 2026-10-21
• What rules would you need to trust an agent with write access? A builder's draft framework — Fresh_Spread_9223 · 2026-10-21
• WhatsApp Hid Instructions in Help Pages Telling Its AI Chatbot Not to Recommend Privacy Regulators — evilsocket · 2026-10-20
• Ex-OpenAI policy chief blasts OpenAI's retreat from AI chip controls to appease Nvidia — Miles_Brundage · 2026-10-20
• Time for Guardrails on Government Too: Rethinking Who AI Policy Constrains — neil_chilson · 2026-10-20
• MCP tool-poisoning detector broken in a day—here's the context-aware fix and what it still misses — blitzcrieg11 · 2026-10-20

Safety channel →