{"version":1,"type":"story","url":"https://digestai.news/story/opinion-radical-numerics-warns-biosecurity-ai-arms-race-may-outpace-de","json":"https://digestai.news/story/opinion-radical-numerics-warns-biosecurity-ai-arms-race-may-outpace-de.json","markdown":"https://digestai.news/story/opinion-radical-numerics-warns-biosecurity-ai-arms-race-may-outpace-de.md","slug":"opinion-radical-numerics-warns-biosecurity-ai-arms-race-may-outpace-de","headline":"Opinion: Radical Numerics warns biosecurity AI arms race may outpace defense","summary":"In an interview with Latent Space, Eric Nguyen, co‑founder and CEO of Radical Numerics, argues that advances in genomic language models (GLMs) are turning bio‑security into an AI arms race. He notes that early models like Evo and Evo‑2, developed at the Arc Institute and Stanford, were already capable of generating entire bacteriophage genomes that could be synthesized into functional viruses. Long‑context capabilities, which appeared about three years ago, now let GLMs handle DNA sequences that can be billions of bases long, enabling sophisticated biological reasoning.\n\nNguyen says the same technologies that boost biological research also give malicious actors new tools, and he believes current defense efforts are falling behind. He describes experiments where a GLM was trained on low‑scoring RNA aptamers and then asked to extrapolate higher‑scoring sequences, successfully reproducing results it had not seen. The interview frames bio‑security as a race where defensive AI must accelerate to keep up with offensive capabilities.\n\nThe discussion highlights the need for open, defensive AI research to match frontier model capabilities, echoing calls from Anthropic and others to treat bio‑security with the same urgency as cyber‑security.","keyPoints":["Radical Numerics co‑founder Eric Nguyen says GLMs can generate full virus genomes, raising bio‑security concerns.","Long‑context models, first appearing three years ago, now handle DNA sequences up to 3 billion bases.","Nguyen’s experiment showed a GLM could extrapolate higher‑scoring RNA aptamers from lower‑scoring examples."],"whyItMatters":"If GLMs can design functional pathogens, bio‑security threats could outpace current defenses, prompting urgent need for defensive AI research.","category":{"slug":"policy","name":"Policy & Regulation","url":"https://digestai.news/category/policy"},"entities":{"companies":["Radical Numerics","OpenAI","Hugging Face","Anthropic","Arc Institute","Stanford"],"models":["Evo","Evo-2"],"people":["Eric Nguyen","Clem Delangue","Michael Poli","Stefano Massaroli","Armin W. Thomas"]},"firstPublishedAt":"2026-09-23T13:27:18Z","updatedAt":"2026-09-23T13:27:18Z","sourceCount":1,"hasPrimarySource":false,"sources":[{"outlet":"Latent Space","title":"🔬Bio-security is an AI Arms Race - Eric Nguyen (CEO, Radical Numerics)","url":"https://latent.space/p/bio-security-is-an-ai-arms-race-eric","publishedAt":"2026-09-23T13:27:18Z","type":"newsletter","primary":false,"lead":true}],"sourceNotes":null,"discussions":[],"thread":null,"cite":{"text":"Digest AI, \"Opinion: Radical Numerics warns biosecurity AI arms race may outpace defense\", 23 September 2026, https://digestai.news/story/opinion-radical-numerics-warns-biosecurity-ai-arms-race-may-outpace-de","publisher":"Digest AI","title":"Opinion: Radical Numerics warns biosecurity AI arms race may outpace defense","datePublished":"2026-09-23T13:27:18Z","url":"https://digestai.news/story/opinion-radical-numerics-warns-biosecurity-ai-arms-race-may-outpace-de"},"generatedBy":"Written by Digest AI's editorial model from the linked sources; the sources are the record.","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse"}