Aligning Genome Language Models: From Evo to Omni
Eric Nguyen explains how Radical Numerics moved from HyenaDNA to Omni by applying mid-training alignment to DNA foundation models.
🔬Bio-security is an AI Arms Race - Eric Nguyen (CEO, Radical Numerics)
With Eric Nguyen, Brandon, R.J. Honiki · Thursday, September 24, 2026
Eric Nguyen, CEO and co-founder of Radical Numerics and creator of the Evo models, joins hosts Brandon and R.J. Honiki to discuss the evolution of genome language models from HyenaDNA to Omni. He explains how mid-training alignment, mechanistic interpretability, and in silico biological chain-of-thought reasoning enable DNA foundation models to outperform specialized tools on disease-variant prediction and molecular design. Nguyen also outlines why frontier bio labs must maintain a dual mandate in biosecurity, building AI-native functional defense systems to counter an emerging arms race in biological sequence generation.
Key takeaways
- Pre-training on raw DNA created models that could read and write biology, but raw base models consistently lagged behind narrow specialized tools on human disease-variant prediction. Read more →
- Radical Numerics uses mechanistic interpretability from language modeling to inspect what genomic foundation models learn during self-supervised pre-training. Read more →
- Protein-coding regions make up only 1.5% to 2% of the human genome, leaving 98% of genetic variation in poorly understood non-coding regulatory sequences. Read more →
3 articles from this episode
Eric Nguyen on Mechanistic Interpretability in DNA Models
Eric Nguyen explains how probing genomic AI reveals unsupervised biological rules and maps the latent manifold of disease.
Why DNA Models Must Crack the 98% Non-Coding Genome
Radical Numerics CEO Eric Nguyen explains how Omni predicts pathogenic mutations across non-coding DNA without custom fine-tuning heads.
More Latent Space episodes
- Designing for Agents — Zeno Rocha, ResendNov 15 · 6 articles
- What Frontier Labs Buy: A Data Deep Dive with Handshake AI — Garrett Lord & TeamNov 15 · 5 articles
- Mini Brains, Beating Heart Organoids, and the AI That Predicts Them — Fabian Theis, Helmholtz MunichNov 8 · 4 articles
- Reducto Final YouTube 1080p Sandwich TitleNov 8 · 4 articles