How to Solve AI Biosecurity by Sophie Kim

·Nuno Sempere··

my best 11,600-word guess—Bioweapon.AI is fi­nally finished!!Some notes from me:Peo­ple who aren’t wor­ried about bioweapons are wrong.The state of biose­cu­rity is re­ally re­ally re­ally re­ally bad and scary and alarm­ing, even in­de­pen­dent of AI.The U.S. has a lot of cheese.Read the full es­say at Bioweapon.AI, which has bet­ter UI than Sub­stack!I may con­tinue edit­ing /​ work­ing on this pro­ject in the fu­ture, though will most likely be tak­ing a break from biose­cu­rity work for a bi...

Read full article →

Related Articles

How many people will attend EAG NYC 2026?
d · Manifold Markets · 9h ago
GPT-6.1+ Astra released in October?
Bayesian · Manifold Markets · 10h ago
October 2026 AI model releases
prismatic · Manifold Markets · 1d ago
Trusted Deployers: Open Model Compromise by Ryan Baker
Ryan Baker · Nuno Sempere · 1d ago
Runtime Assurance Meets Corrigibility: What Happens When an Independent Safety Boundary Blocks an AI Agent? by Idorenyin
Idorenyin · Nuno Sempere · 2d ago