Insiders push back on AI-virus threat models: ordering viral fragments as a rando gets you reported to the FBI
basedjensen · x · 2026-09-15
Responding to Esvelt's SecureBio stunt and analyses from IFP and similar policy shops, the poster argues their threat models are incoherent: a lone wolf smart enough to secretly bootstrap a real virus from DNA fragments but too dumb to synthesize oligos doesn't make sense.
Core points: reputable synthesis shops follow voluntary screening protocols, and a non-institutional individual ordering viral fragments risks an FBI report; the practical defense already exists even as laws tighten. Phosphoramidite oligo synthesis requires basic orgo skills (water-sensitive chemistry) but isn't that hard, and 40mers suffice to tile genes — DNA synthesis shops aren't silicon fabs, so they don't form a hard barrier.
More from AGI Musings
- "Those paying least attention are most certain": a Don't Look Up analogy for AI risk — callme_e · 2026-09-15
- davidmanheim: prosaic biorisks justify billions in biodefense, but won't stop ASI — davidmanheim · 2026-09-15
- David Sacks mocks AI doom predictions as another 'engineered crisis' — beffjezos · 2026-09-15
- MIRI's comms strategy: convince world powers to shut down frontier AI before it's too late — davidmanheim · 2026-09-15
- Zena Hitz Warns LLMs Obscure the Primary Sources of Human Learning — AnnaCiaunica · 2026-09-15
- Analytic philosophy veteran: most takes on philosophy of mind are confidently wrong — danbri · 2026-09-15