Liam Fedus
person on 6 shows · 5 statements across 2 episodes · said 7 times in 6 episodes since 2025
No Priors 2 
Catalyst 2 
Latent Space 1 
TBPN 1 
20VC 1 
the a16z Podcast
On the record as a speaker too: Liam Fedus across every show, appearances and statements → this page counts the times other people say the name.
Mentions by year, every show
tap a year for its mentions
Catalyst 2No Priors 220VC 1Latent Space 1TBPN 1
2026 4 mentions in 4 episodes 1 per episode
- Why AI Labs With Unlimited GPUs Still Fail — Anjney Midha, AMP
- FULL INTERVIEW: Anjney Midha on Fixing AI’s Biggest Bottleneck
- The Early Days of Anthropic & How 21 of 22 VCs Rejected It | The Four Bottlenecks in AI | Anj Midha · 20VC with Harry Stebbings
- Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI
- every mention in 2026, scene by scene →
2025 3 mentions in 2 episodes 2 per episode
every mention on every show, scene by scene, with the transcript →
5 statements about Liam Fedus, every show
Fedus: Post-Higgs experimental bottlenecks drove physicists to transition into AI
“After the discovery of the Higgs I think a lot of high energy physicists were sort of looking for what's next. Ultimately it becomes bottlenecked on the new apparatus for, you know, pushing the next energy frontier. And I think a lot of physicists were looking…”
Fedus: Hybrid automated lab systems already produce large amounts of reliable data
“Already the reliability of the sort of like hybrid systems is sufficient to produce Huge amounts of reliable data”
Fedus: Periodic Labs uses off-the-shelf robotics instead of building custom hardware
“Right now we're using almost like more like off the shelf robotics. It's like very simple, very commoditized not doing like a huge amount of innovation on, on that front.”
Fedus: Pre-training on domain data outperforms retrieval-augmented generation
“However, as we've seen with things like ChatGPT and other things, when you pre-train on the data, when you actually encode the knowledge into the weights, it's not just a retrieval system, you have a richer, deeper understanding of the material.”
Fedus: High-compute reinforcement learning is essential for AI tool use
“High compute reinforcement learning is really effective. This is how you should think about the strategies it's using. This is how you create effective tool using towards those problems, and this is how you optimize it effectively.”