Create

Sign in to ReadmeX

or

How we handle your data: Privacy Policy

ReadmeX

Your AI reading companion

What would you like to catch up on?

Pick a briefing, or ask your own question.

Your community and people briefings stay personal to you.

Story

Periodic Labs on 'synthesis superintelligence': AI that experiments in the physical world

AI summary

On the Latent Space podcast, Periodic Labs' Liam Fedus and Ekin Dogus Cubuk described the lab's vision of "synthesis superintelligence": reinforcement learning grounded in physical experiments, combined with AI-powered materials characterization, simulation and density functional theory, and high-throughput automated labs. They argue that scientific discovery differs fundamentally from math and coding because it requires reasoning under noise, uncertainty and missing information, and that failed experiments may be among the most valuable training data. The episode also covers giving every piece of lab equipment "140 IQ," compressing decades of trial-and-error into months through autonomous experimentation, and why even future frontier models will still need to run real experiments.

In-depth insight

The interview lays out a route different from simply training on more internet data: letting models experiment in the real world and learn from the entire process of doing science, not just published results.

It’s hard to believe that Periodic was only launched last September:

One year later, it is considered one of the pre-eminent AI scientist labs, with dizzying talent density and astonishing progress in the autonomous lab buildout:

Most people are familiar with the standard credentials of Liam and Dogus, but we found an incredible “talent slope” while learning more about Periodic, where each successive employee seems more impressive than the last:

From building AI systems that reason over noisy physical experiments to creating laboratories where every instrument can become intelligent, Periodic Labs is betting that the next frontier of AI won’t come from simply training on more internet data, it will come from letting models experiment with the real world. In this episode, Periodic Labs’ Liam Fedus and Ekin Dogus Cubuk join swyx and Brandon to explain why scientific discovery is fundamentally different from math and coding, and what it takes to build AI scientists that can actually discover new materials.

We go deep on Periodic’s vision for “synthesis superintelligence”: reinforcement learning grounded in physical experiments, AI-powered materials characterization, simulations and density functional theory, high-throughput labs, and systems that learn from the entire process of doing science rather than only its published results. Liam and Dogus also explain why frontier models will still need experiments, why failed experiments may be some of the most valuable training data, what it means to give every piece of lab equipment “140 IQ,” and how autonomous experimentation could compress decades of scientific trial-and-error into months.

Some source content is omitted. Continue reading on the original site.

#Periodic Labs#Liam Fedus#Ekin Dogus Cubuk#Latent Space

Original · 38 min readRead the original
Group chat

No comments yet. Start the conversation.