scripod.com
AI for Atoms: How Periodic Labs is Revolutionizing Materials Engineering with Co-Founder Liam Fedus

Overview

Shownote

Highlights

Transcript

Chapters

Pins

AI for Atoms: How Periodic Labs is Revolutionizing Materials Engineering with Co-Founder Liam Fedus

No Priors: Artificial Intelligence | Technology | Startups

Apr 03
No Priors: Artificial Intelligence | Technology | Startups

No Priors: Artificial Intelligence | Technology | Startups

Apr 03
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

What happens when you apply the scaling laws of large language models to the physical work of atoms? Elad Gil sits down with Liam Fedus, co-founder at Periodic Labs, which is pioneering an AI foundation lab for atoms. Liam discusses how he pivoted from dar...

Highlights

Liam Fedus, co-creator of ChatGPT and former VP of post-training at OpenAI, discusses his new venture Periodic Labs, which applies AI scaling laws to materials science and chemistry. The conversation explores how large language models can be connected to the physical world to accelerate scientific discovery and overcome data bottlenecks in material science.
00:05
An AI foundation lab for atoms
00:39
Physics provides a principled approach to problem-solving.
05:14
Connecting AI to the physical world is essential
09:29
It is an active loop, not a static data pool.
09:52
Generalization works for quantum systems but not fluid dynamics
11:31
Language models orchestrate experiments and coordinate specialized neural nets.
15:29
AI education reaches every young girl
16:10
AI-driven atomic rearrangement synthesis could match digital progress.
17:45
Multidisciplinary collaboration drives scientific progress
19:41
Compute costs dominate over physical infrastructure
21:12
Periodic Labs is hiring for AI and physical science roles
21:44
Intelligence is not a scalar.
23:30
Software engineering self-improvement is happening now
27:59
Reliable general robotic systems are needed to scale high-throughput experiments.
27:59
AI interfacing with the physical world is a massive opportunity

Chapters

Liam Fedus Introduction
00:00
Liam’s Background at Google Brain, OpenAI
00:39
From ChatGPT to Materials and Atoms
05:14
Training Data in the Physical World
06:34
Generalization Across Domains
09:52
Models as an Orchestration Layer
11:31
Commercialization and Business Model
12:48
How Periodic’s Success May Shape the Future
16:10
Multidisciplinary Scaling
17:45
Capital and Compute
19:41
Hiring at Periodic
21:12
Thoughts on AGI and ASI
21:44
Timeline for Machine-Directed Self-Improvement
23:30
Automation and Data Generation
25:39
Why Liam is Excited About the Future of Robotics
27:59

Transcript

Elad Gil: To No Priors, we're talking with Liam Fedus. Liam is one of the co-creators of ChatGPT, which I think almost everybody uses at this point. He was the VP of post-training at OpenAI, and before that was at Google Brain, where he worked on a variety...