Liam Fedus

Building industrial-scale science at @periodiclabs

Past: VP of Post-Training @OpenAI; Google Brain

[Liam Fedus @LiamFedus]

We built high-throughput materials labs in Menlo Park to create a loop between experiments and models. The labs generate fresh data, the models learn from it, and then help us decide what to try next.

Using only 1,300 H200s, plus months of our experimental data, we mid-trained and RL’d an open-source model to surpass GPT-6 Astra on our analysis benchmark. We call it Neon.

This is real footage from our lab. We’re focusing first on hard problems in materials science, including superconductors, magnets, and semiconductor materials.

Read our blog posts below.

[SGLang @sgl_project]

@LiamFedus is joining us at SGLang Summit for a fireside chat on Emerging AI Labs!

Liam helped create ChatGPT and led post-training at OpenAI. Now at @periodiclabs, he’s building AI scientists that learn from real-world experiments in autonomous labs. Join us on Nov 13 to hear his perspective on the next wave of AI labs and the opportunity to turn AI advances into scientific discoveries.

🎙 Nov 13 · Fireside Chat · Fort Mason, SF
🔗 https://t.co/9iOHh4iycr
Expo Pass holders can watch main stage sessions via livestream onsite.

[Liam Fedus @LiamFedus]

@xanderai This reminds me of when @vwxyzjn one-shotted a @wandb replacement

[Liam Fedus @LiamFedus]

This is a fun era to build a startup

[Xander Dunn @xanderai]

Our spend on DataDog is too high ->

Hey Opus 5.5 switch everything to Grafana and VictoriaMetrics ->

1 day later we spend 38x less on all logs, alerts, and graphs. No loss of functionality.

[Liam Fedus @LiamFedus]

PageRank for startups, @periodiclabs coming in at #5

[Deedy @deedydas]

Top 30 Startups and BigTech companies based purely on ~50,000 engineer job changes in the last 3 years.

If engineers leave a company to join another, it ranks higher especially if the company they come from ranks highly.

Here are the final results:

[Liam Fedus @LiamFedus]

In the long run, put optimization pressure on AI for unbounded opportunities, e.g., science and tech (which build upon prior art) rather than accounting (bounded once an AI successfully follows all tax laws and regulations).

[TBPN @tbpn]

.@dylan522p explains his position that the first trillion dollars of AI spend won't be used to replace white-collar work:

"Replacing white-collar work is something that will happen, but that's not going to drive the first trillion dollars of AI spend. Those are not industries that are going to diffuse AI capabilities quickly. It's going to take them quite some time to reorder how their OpEx is spent. Because companies have a lot of inertia with regard to people's jobs, changing how their compensation works, how many people work on what, etc."

"There are areas where we think that intelligence is very, very important, though. These include areas like chip design, cybersecurity, financial services, life sciences, materials science, software engineering. You can absorb unlimited intelligence [in these areas]. And there are two reasons for that."

"One is that it's a domain where humans aren't anywhere close to the frontier and there's just so much to discover. Think materials science, chip design, anything in the scientific realm. We figure out new stuff, we deploy it, we make everyone's lives better."

"And there's the other world which is adversarial. Specifically, finance and cybersecurity. It's human vs. human, and we're constantly going to be leveling up, either trying to arb each other like quants do, or trying to hack each other and then play defense. So there's this constant demand for intelligence because when you have more intelligence than the other parties in the market, you win."

[Liam Fedus @LiamFedus]

Catch a high-alpha talk by @hsu_byron on our collaboration between Periodic and RadixArk on building our inference stack

[Byron Hsu @hsu_byron]

Periodic Labs is heading to #SGLangSummit 2026!

We’re building AI scientists and autonomous labs that learn from real-world experiments. This requires powerful models, robust infrastructure, and a tight feedback loop between ideas and experimental results.

I’ll be talking about how we’re building RL infrastructure for physical experiments—from developing custom trainers and sandboxes from scratch to adapting SGLang to optimize RL workloads.

Excited to see you all there! @sgl_project @radixark

[Liam Fedus @LiamFedus]

@ypatil125 Congrats, Yash and team! Incredible progress

[SGLang @sgl_project]

Proud to welcome our newest sponsors to SGLang Summit.

In just one week since our public announcement, our sponsor lineup has grown from 11 to 36. Thank you to every sponsor helping bring this community together.

Join 2,000+ engineers, researchers, and builders from across the AI ecosystem for technical sessions, hands-on workshops, and the GPU Expo.

Hear from Lip-Bu Tan, Lilian Weng, Soumith Chintala, Aravind Srinivas, and more. Meet the teams building across AI hardware, infrastructure, models, and applications.

🗓️ November 12–13, 2026
🌉 Fort Mason, San Francisco
🎟️ Get your tickets: https://t.co/7C6TzrkJYV

[Liam Fedus @LiamFedus]

Mirror Mirror built the licensing layer that lets people grant brands permission to use their likeness for AI generation. Be a model!

Now brands can digitally license with real people instead of using generic fake AI faces or deepfakes.

And there's no better person than @yusan_lin to build this. Her career has spanned both sides as a model and an AI researcher.

[Yusan Lin @yusan_lin]

AI slop isn't just tacky and uninspired. It can get you sued.

Generating fake AI faces that resemble real people, or using someone's likeness without permission, is creating new legal and reputational risks for brands.

So we built @mirrormirror_ai as part of a16z @speedrun, a marketplace where brands can digitally book real models and creators to generate AI content using their likeness, with their permission.

Brands browse and book identity-verified talent, agree on usage rights, and generate content within those terms. Every approved image comes with a signed license and embedded provenance.

We're seeing brands produce lookbooks and campaign visuals up to 10x faster, with up to 90% cost savings. Meanwhile, models and creators can earn additional income from their digital likeness without having to show up for another shoot.

Everyone is racing to make AI generation faster, cheaper, and more realistic. But there's still a huge missing piece: how do you license the rights to the people appearing in AI-generated content at scale?

That's why I'm excited about what we're building. Not just another AI generator, but the licensing and rights infrastructure for AI-generated content.

https://t.co/7u72P5xvmq

[Liam Fedus @LiamFedus]

@patchwanders Yep! Though we did have a few very active folks. But it wasn’t mainstream

[Liam Fedus @LiamFedus]

ChatGPT’s success was thoroughly surprising (at least to me).

GPT-3.5, the backbone for ChatGPT, was already out for months ahead as a completion API, our internal usage didn’t have particularly high usage, and there was no commercially successful chatbot until that point.

Always enjoy my chats with @BrendanFoody. This one covers ChatGPT and my experiences across physics, ML and the intersection of both.

[Brendan (can/do) @BrendanFoody]

1/ ChatGPT had the internet. AI for science has no equivalent.

@LiamFedus helped create ChatGPT and went on to lead post-training at OpenAI. Now he’s building Periodic Labs because better models alone won’t get us to an AI scientist.

We talked about why Periodic is building physical labs and what Liam learned from being inside OpenAI as ChatGPT took off on today's episode of The Frontier.

[Liam Fedus @LiamFedus]

@DBahdanau is 🐐

Over the years, I've had the awesome fortune of working with the best ML researchers in the field, plus the incredible new cohort of future leaders who IC the AI tech we use daily.

And Dima still stands out as a rare trifecta: great leader x brilliant researcher x exceptional systems engineer.

[rohan anil @_arohan_]

Bahdanau is 🐐

He cooked edge transformers years before we got to 2-simplicial as well.

[Rohan Pandey @khoomeik]

yeah this checks out

my first few months at periodic, the weekly midtraining meeting was:
1. guy who trained the first trillion param LLM
2. guy who invented the attention mechanism
3. me

you learn very quickly surrounded by this density of talent. we're hiring btw.

[Rishabh Agarwal @agarwl_]

If you are interested in working on science of scaling RL, please apply!
You'll be closely working with me

[Liam Fedus @LiamFedus]

Decision-making under uncertainty and partial observability is harder for AI, but in the long run, it will ultimately be more impactful.

We talk about taking those steps beyond the comfortable world of games, code, and math.

[Mario Gabriele @mariogabriele]

"A system that can literally engineer matter is going to be of key importance to everybody."

@LiamFedus and @ekindogus are the co-founders of @periodiclabs, a San Francisco-based startup building AI systems and autonomous laboratories to discover new materials, starting with a higher-temperature superconductor. Fedus previously helped build ChatGPT and ran post-training at OpenAI; Cubuk led materials and chemistry research at Google DeepMind.

(3:14) What it takes to build a synthetic superintelligence
(16:44) Dogus and Liam's path to AI
(23:28) Lessons from launching ChatGPT
(35:10) What AI learns across experiments
(49:23) The difference in LLM performance on math vs. science
(54:09) The competitive landscape
(55:39) Final meditations

[Liam Fedus @LiamFedus]

With Periodic Neon and Gemini Argon now taken, the race is afoot for Krypton

Congrats, Google and the team, on the impressive model and rejoining the frontier! The ongoing intelligence explosion is creating massive consumer surplus for users and startups. PhD on demand, pennies to run.

[Sundar Pichai @sundarpichai]

Lots of discussion out there about our next model(!), so I wanted to give an early look as soon as possible. Introducing Gemini 4 Argon!

It shows frontier performance in complex workflows, cyber defense and software engineering. Teams are using it extensively at Google, from coding to quantum computing, great feedback.

Here’s a look at the benchmarks:

[Liam Fedus @LiamFedus]

Periodic 🤝 SGLang to further improve open-source software

[Byron Hsu @hsu_byron]

Happy to contribute a fast and correct RL sampling mask from @periodiclabs, developed jointly with @nanjiangwill @agarwl_ and now running in our production RL pipeline!

[Jason Kelly @jrkelly]

💯All roads for AI science lead to real labs.

Most of science is observation-limited! AI will need to run experiments if we want to learn faster in the world of atoms.

[MTS @MTSlive]

SITUATION DETECTED: Anthropic has set up a wet lab in the San Francisco Bay Area for physical biology work. It wants to unlock treatments for rare diseases, and its research has now gone beyond in silico evaluations, per Reuters.