Welcome, humans.

Most AI models write text, images, or code. Eric Nguyen helped build one that can read and write DNA.

This is a big deal, especially after the reality that AI can now design viruses went viral last week (pun very much intended).

In our latest podcast episode, Corey and Grant spoke with Eric Nguyen, co-founder and CEO of Radical Numerics, about AI-designed CRISPR systems, complete viral genomes, personalized medicine, and what happens when biological AI can create things nature never did.

Watch and/or listen now: YouTube | Spotify | Apple Podcasts

Eric’s earlier model, Evo, helped scientists design working CRISPR-Cas systems and later generate a complete bacteriophage genome.

His new company is pushing toward multimodal biological models that combine DNA with RNA, proteins, epigenetics, and other signals. In other words (our interpretation here), they’re trying to model the human body and solve all illness.

Here’s our favorite parts:

• (09:41) AI generated a complete viral genome: Eric explains how scientists used Evo to create the full DNA sequence for a bacteriophage, then why that same capability creates an obvious biosecurity problem.

• (31:05) Two years of experiments, reproduced in 30 minutes: A new Radical Numerics teammate used Omnii to rank Alzheimer’s-related genes in roughly half an hour after scientists spent two years validating them in the lab.

• (38:39) What is a “deepfake virus”? Eric explains how AI could change the genetic spelling of a virus while preserving its function, potentially bypassing sequence-based detection.

• (44:33) Custom biological countermeasures on demand: The group explores AI-designed bacteriophages that could target drug-resistant bacteria when ordinary antibiotics fail.

• (45:24) Personalized medicine at the limit: Could a future doctor take a snapshot of your biology and design a bespoke treatment for you? Eric thinks AI may finally make that old promise practical.

The biggest idea is that biological AI is moving from reading life to designing it. That could mean better medicines and gene therapies, but it also creates a new security problem: the same models that can design novel biology may be needed to detect it.

Why watch this? Because Eric makes genomic AI understandable without sanding off the weirdness. You will see how DNA becomes a language-model problem, why biology needs enormous context windows, and where the line sits between medicine and misuse.

Watch / listen now: YouTube | Spotify | Apple Podcasts

P.S. For the philosophical dessert, jump to 49:10, when Grant asks a genomic-AI researcher for his best theory on the origin of life.

Keep scrolling for the security problem hiding inside AI agents, a deeper look at biological AI, and four recent Neuron conversations worth watching next.

THIS EPISODE WAS BROUGHT TO YOU BY…

Least Privilege for the Age of AI Agents

AI agents are growing 40% year-over-year inside enterprises. 7% of orgs already had an agent-related security incident this year.

The problem: AI agents don't create new permissions — they weaponize the ones already there. Same cloud keys, same tokens. No scoping. No expiration. No one watching.

BeyondTrust AI Agent Security:

→ Sees every agent — including shadow AI

→ Attributes every action to human or agent

→ Blocks risky commands, enforces approval before agents act

→ Works across Claude Code, Copilot, Cursor, more — one policy, any vendor

🔴 LIVE TOMORROW: Learn AI Video Prompting with the LTX Team

Thursday, August 13 at 10 AM PT: We’re going LIVE with the team behind the newly announced LTX-2.5 for a total beginner’s guide to AI video generation and video prompting.

We’ll go hands-on with the new model while the LTX team teaches us how to write better video prompts, control shots and motion, describe scenes clearly, and understand what video models actually pick up from your instructions.

No AI video experience required. Bring your weird prompts, difficult questions, and video ideas. We’ll demo LTX-2.5 live, put it through its paces, and answer questions with the team that built it.

Read the LTX-2.5 announcement | View it on Hugging Face

DNA is a sequence of just four letters, but a human genome contains roughly 3B base pairs. Eric’s team treats those sequences like language so models can learn biological grammar from raw data.

• Evo: a genome language model that learned from DNA across the tree of life and helped generate CRISPR-Cas systems and a bacteriophage genome.

• Omnii: Radical Numerics’ newer direction, combining DNA with RNA, proteins, epigenetics, and other biological signals.

• The goal: understand how biological changes cascade through a system so researchers can design safer therapies and respond faster to new threats.

The hard part is validation. A model can propose a sequence quickly, but scientists still need physical experiments, manufacturing, and safety checks before biology leaves the computer.

Explore the work:

• Visit Radical Numerics

• Read the Omnii health preview

• Read the original Evo research

• Read the Evo 2 paper

🎙️ In Case You Missed It…

1. Want to run powerful AI without sending everything to the cloud?

TL;DW: Intel’s Dr. Olena Zhu explains why the future of AI may be hybrid: private and repetitive work stays local, while harder reasoning gets routed to bigger cloud models. Plus, she shares a staggering fact: if current trends hold, we might have Fable-class AI on our (powerful) laptops “within two years.”

Why you should watch: It turns “local AI” from a privacy slogan into a practical architecture for agents, cost, and reliability, along with the tools you can use to do it.

Watch / Listen: YouTube | Spotify | Apple Podcasts

2. Building something with AI? Watch: AWS Put a CTO Inside Claude Code

TL;DW: AWS startup leader Deap Ubhi explains how AI compressed startup iteration from months into days, while security, infrastructure, and reliability still separate a prototype from a business.

Why you should watch: It shows when builders should move fast and when technical shortcuts become expensive traps.

Watch / Listen: YouTube | Spotify | Apple Podcasts

3. How do you make truly autonomous surgery trustworthy? This interview will teach you…

TL;DW: Mathias Unberath explains why autonomous surgery is difficult, how developers test rare failures, and what reliability means when mistakes have physical consequences.

Why you should watch: It is a sharp guide to the gap between a technical demo and a dependable real-world system.

Watch / Listen: YouTube | Spotify | Apple Podcasts

4. Want to open AI’s black box?

TL;DW: Goodfire CEO Eric Ho explains features, circuits, confidence signals, and how researchers may inspect what models are doing internally.

Why you should watch: It replaces “the model is magic” with a practical look at debugging and steering AI systems.

Watch / Listen: YouTube | Spotify | Apple Podcasts

New episodes of The Neuron: AI Explained explore the breakthroughs, businesses, and people shaping artificial intelligence.

Subscribe on YouTube by clicking below to help us get even more amazing guests like this one to teach you new things about AI every week!

Stay curious,

The Neuron Team

P.P.S: Love the newsletter, but don’t want these podcast announcement emails? Don’t unsubscribe. Adjust your preferences to opt out of them here instead.