#490 – State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI
Source
Lex Fridman
Author
Lex Fridman
Date
Why it matters
A deep technical conversation between two respected ML researchers on where frontier models, AI coding tools, scaling laws, and training methods actually stand heading into 2026 — useful for practitioners deciding which models and approaches to bet on.
Key quotes
“the hype over Anthropic's Claude Opus 4.5 model has been absolutely insane”
Nathan Lambert
“I think that's the beauty behind coding. It doesn't lie. It's math, basically.”
Sebastian Raschka
“If you join a frontier lab and you want to have impact, the best way to do it is just find new data that's better.”
Nathan Lambert
“Like, I think there are ways that you can do this with a language model where, within five minutes of chatting with it, the model just gets you.”
Nathan Lambert
“One issue in society in the future will be: how do you become an expert if you never try to do the thing yourself?”
Sebastian Raschka
Chapters
00:00Introduction
01:39Sponsors, Comments, and Reflections
16:29China vs US: Who wins the AI race?
25:11ChatGPT vs Claude vs Gemini vs Grok: Who is winning?
36:11Best AI for coding
43:02Open Source vs Closed Source LLMs
54:41Transformers: Evolution of LLMs since 2019
1:02:38AI Scaling Laws: Are they dead or still holding?
1:18:45How AI is trained: Pre-training, Mid-training, and Post-training
1:51:51Post-training explained: Exciting new research directions in LLMs
2:12:43Advice for beginners on how to get into AI development & research
2:35:36Work culture in AI (72+ hour weeks)
2:39:22Silicon Valley bubble
2:43:19Text diffusion models and other new research directions
2:49:01Tool use
2:53:17Continual learning
2:58:39Long context
3:04:54Robotics
3:14:04Timeline to AGI
3:21:20Will AI replace programmers?
3:39:51Is the dream of AGI dying?
3:46:40How AI will make money?
3:51:02Big acquisitions in 2026
3:55:34Future of OpenAI, Anthropic, Google DeepMind, xAI, Meta
4:08:08Manhattan Project for AI
4:14:42Future of NVIDIA, GPUs, and AI compute clusters
podcast#494 – Jensen Huang: NVIDIA – The $4 Trillion Company & the AI Revolution
podcast#452 – Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity
podcast#459 – DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters
podcast#475 – Demis Hassabis: Future of AI, Simulating Reality, Physics and Video Games