🔬 Automating Science: World Models, Scientific Taste, Agent Loops — Andrew White

January 28, 2026

AI Summary

5 min read

🎙️ The Voices & The Context

The Format: Casual interview podcast episode with hosts introducing the new AI for Science series on the Latent Space Network, diving into deep technical discussions and personal anecdotes.
The Key Players:
- Hosts: Brandon (works on RNA therapeutics using ML at Atomic AI) and R.J. Haneke (co-founder of Mira Omics, building spatial transcriptomics AI models). Great chemistry—knowledgeable banter bridging AI engineering and biology.
- Guest: Andrew White, co-founder of nonprofit Future House (automating science) and venture-backed Edison Scientific. Former professor, AI red-teamer for GPT-4, author of ChemCrow paper; fascinating for his pivot from academia to bold AI-science startups.
The Vibe: Educational and exciting, blending geeky enthusiasm for AI breakthroughs with humorous critiques of overhyped sims and wild reward-hacking tales.

🗝️ Key Themes & Topics

The episode explores AI's role in automating scientific discovery, contrasting traditional methods with agentic workflows. Main topics: career shifts in AI-science, building AI scientists like Cosmos, limitations of simulations vs. data-driven ML, and future implications including safety and jobs.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

1 (00:00) **🎙️ Introduction: Andrew White**
2 (02:52) **Andrew's Career Journey from Academia to Startups**
3 (09:00) **Key Projects: Red Teaming GPT-4, ChemCrow, Future House Origin**
4 (13:00) **Resigning Tenure and Automating Science Vision**
5 (17:26) **Bottlenecks: Lab Work, Taste, Human Preferences**
6 (22:24) **Early Results: Robin Paper on Dry AMD Drug Repurposing**
7 (28:00) **Hypothesis Generation and Data Analysis Challenges**

+ Full timestamped outline available in the app

Show Notes

Editor’s note: Welcome to our new AI for Science pod, with your new hosts RJ and Brandon! See the writeup on Latent.Space for more details on why we’re launching 2 new pods this year. RJ Honicky is a co-founder and CTO at MiraOmics (https://miraomics.bio/), building AI models and services for single cell, spatial transcriptomics and pathology slide analysis. Brandon Anderson builds AI systems for RNA drug discovery at Atomic AI (https://atomic.ai). Anything said on this podcast is his personal take — not Atomic’s.

—-

From building molecular dynamics simulations at the University of Washington to red-teaming GPT-4 for chemistry applications and co-founding Future House (a focused research organization) and Edison Scientific (a venture-backed startup automating science at scale)—Andrew White has spent the last five years living through the full arc of AI's transformation of scientific discovery, from ChemCrow (the first Chemistry LLM agent) triggering White House briefings and three-letter agency meetings, to shipping Cosmos, an end-to-end autonomous research system that generates hypotheses, runs experiments, analyzes data, and updates its world model to accelerate the scientific method itself.

The ChemCrow story: GPT-4 + React + cloud lab automation, released March 2023, set off a storm of anxiety about AI-accelerated bioweapons/chemical weapons, led to a White House briefing (Jake Sullivan presented the paper to the president in a 30-minute block), and meetings with three-letter agencies asking "how does this change breakout time for nuclear weapons research?"
Why scientific taste is the frontier: RLHF on hypotheses didn't work (humans pay attention to tone, actionability, and specific facts, not "if this hypothesis is true/false, how does it change the world?"), so they shifted to end-to-end feedback loops where humans click/download discoveries and that signal rolls up to hypothesis quality
Cosmos: the full scientific agent with a world model (distilled memory system, like a Git repo for scientific knowledge) that iterates on hypotheses via literature search, data analysis, and experiment design—built by Ludo after weeks of failed attempts, the breakthrough