🔬 The Lab of the Future Should Feel Like a Data Center — Andy Beam & Rafa Gómez-Bombarelli, Lila Sciences
July 16, 2026
AI Summary
5 min readLila Sciences is building what it calls an "AI science factory"—a physical laboratory designed to operate like a data center, where rows of instruments generate experimental data at machine speed rather than human speed. The core bet is that science can serve as an infinite token generator for training AI models, using nature and experiments as a verifier in a reinforcement learning loop. As CTO Andy Beam puts it, "the lab of the future should feel like a data center, rows of server racks as densely packed as possible."
The Bitter Lesson Applied to Science
The company's thesis is a direct application of the "bitter lesson" of AI: methods that scale and generalize beat those that don't. Large language models succeeded because of internet-scale data, but that data is now exhausted. Beam notes, "as Elia said at NeurIPS last year, we have but one internet. It's the fossil fuel. We fracked. We got every ounce of data that we could out of the internet, but it's gone."
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
Never miss an episode of Latent Space: The AI Engineer Podcast
Get every new episode summarized in your inbox — free, ~5 minutes to read.
No spam. Unsubscribe anytime.
What you'll learn
- 1 (00:00) **The Bitter Lesson Applied to Science** - Lila is all-in on scale: methods that scale beat those that don't, and the next frontier is using science as an infinite data generator for post-training.
- 2 (01:00) **Guest Introductions: Andy Beam & Rafa Gómez-Bombarelli** - Andy (CTO, ex-Generate Biomedicines, Harvard) and Rafa (CSO Physical Sciences, co-founder, MIT) describe their paths from early deep learning for science to founding Lila.
- 3 (05:40) **Lila's Thesis: Science as an Infinite Token Generator** - The core idea: use nature and experiments as verifiers for RL-based post-training at scale, pushing the frontier of reasoning models.
- 4 (07:44) **The Runtime of Data Collection** - Different experiments have different feedback timescales (days to weeks); the key is multiplexing and synchronizing training across these timescales.
- 5 (10:06) **The Lab as a PCI Bus** - The physical lab is a flexible graph of instruments connected by a magnetic levitation transport layer, allowing reconfiguration for new experiments.
- 6 (13:28) **AI-Generated Experimental Designs** - The model can design novel protocols zero-shot (80% success vs. 0% for humans), but fully free-form experimentation is the goal.
- 7 (18:30) **Scientific Rigor and Verification** - AI science must be held to the same standards as human-led science; the platform enables rapid verification and explanation of experimental outcomes.
+ Full timestamped outline available in the app
Show Notes
Imagine a dark warehouse. Racks and racks of devices with wires, tubes, and electronics sticking out. The next AI data center? No. This is Lila Sciences‘ dream for the future of science. A dark warehouse full of AI-guided robotics and lab equipment, cranking out new experiments 24/7, building toward a scientific superintelligence.
Their automated lab is almost hypnotizing to watch. They have floating plates zipping around on Wall-E-esque tracks, used vision-language models to control Windows 95 boxes, and created the world’s largest collection of voided warranties. In the process they’ve built a massive library of scientific reasoning tokens. Over 10 trillion of them, all experimentally validated.
No warranties were voided in the making of this video
To say Lila is ambitious is an understatement. Their goal is a scientific superintelligence wired directly into the wet lab. They are all in on the bitter lesson, and the thesis follows from it: a lab is an infinite token generator. Produce data at scale, and the synergies give you a general reasoner that can tackle any scientific problem. They are committing hard. Biology, chemistry, drug discovery, and materials science, all at the same time. Time will tell if it works, but it is an exciting hypothesis.
In our latest episode we sat down with Lila’s very own Andy Beam (CTO) and Rafa Gómez-Bombarelli (CSO, physical sciences) and went on a journey through the possibilities of AI-run science, almost as wide-ranging as Lila’s goals.
Did we mention they do both materials science and biology? In the same AI science factory? Same time, same lab, same AI. Finally a guest who can settle a long-running debate we’ve had amongst ourselves: is biology or materials science harder?
Watch to find out!
We discuss:
* The internet is spent, science is next. Why Lila thinks the scientific method is the last untapped internet-scale dataset, and why they treat RL as a data generation mechanism with nature as the verifier.
* The lab as a data center. Instruments as nodes on a graph, a magnetically levitating “PCI bus” transport layer between them, orchestration as a slurm queue. Andy is not short on analogies.
* Why Lila insists it is not an automation company. They optimize for flexibility and generalizability over raw throughput, which means humans stay below the API line wherever automating does not pay.
* Your experiment has a runtime. We put More from this podcast