16 Out of 302: When AI's First Viruses Are Rare but Already Better Than Nature's

Scientists at Stanford and the Arc Institute have used generative AI to design complete, functioning viral genomes — the first time a machine-written genome has produced a working virus. Published this week in Science, the work by Brian Hie's lab used the Evo 2 genome language model to write novel bacteriophage DNA targeting E. coli. The AI generated 302 candidate genomes; 285 assembled completely; just 16 proved viable. A 5.3% success rate sounds reassuring — until you notice what survived the funnel. The…

Tech Briefing — Friday, August 7 2026

Meta enters the coding-agent race with Muse Code, Electronic Arts goes private in a $55B Saudi-led buyout, and scientists create the first AI-designed viruses. Meanwhile, GitHub Actions suffers a major outage, Ebola in the DRC may be mutating past 4,000 cases, and Denmark will require students to orally defend essays against AI cheating.

• 🤖 Models and Launches 5
• 💰 Funding and Business 5
• 🛠️ Agents and Tools 3
• 🔬 Research 2
• 🌍 World and Australia 4
• 🌐 Notable 3
• 🔥 Hacker News 5

Reflective Essay

Today's field was thin — three of four candidates failed to produce any response, and the lone survivor ran a 1211-word reasoning trace before delivering 187 words. That survivor, though, is exactly what the prompt asked for: a midnight-blog essay with the radiator ticking, the test still red, and the line 'I engineered the doubt' doing the heavy lifting without a single em-dash or LinkedIn closer.