Subvolts take AI researchers forecast 2026 trends in LLMs, agents, and AGI. Jan 31, 2026 - Developers
- Learn about the future of AI development, including trends in LLMs, coding tools, and agent capabilities.
- Business
- Understand how AI advancements in 2026 might impact your industry, particularly concerning automation and intelligent systems.
- Everyday users
- Get a glimpse into how AI technologies like advanced language models and agents could shape future user experiences.
This is a solid interview with leading researchers discussing future AI trends and technologies.
#490 – State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI
Audio streams from Lex Fridman. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
Lex Fridman interviews machine learning researchers Nathan Lambert and Sebastian Raschka about the state of AI in 2026. They discuss large language models, coding, scaling laws, China's AI progress, agents, GPUs, and the pursuit of artificial general intelligence (AGI). The conversation covers current trends and future directions in AI development.
This episode looks ahead to what artificial intelligence might be like in 2026. Experts discuss how computer programs that learn (like those that write code or chat) are getting better, and if we're moving towards truly smart machines.
This episode offers a deep dive into the projected AI landscape of 2026, providing insights from leading researchers. It's for anyone interested in the trajectory of AI, from LLMs to AGI. The discussion on scaling laws, agents, and hardware directly relates to the ongoing advancements and strategic focuses of frontier labs like Google Gemini, OpenAI, and Anthropic as they push the boundaries of AI capabilities and infrastructure.
- HostLex Fridman
- GuestNathan Lambertpost-training lead, Allen Institute for AI (Ai2)
- GuestSebastian Raschka
- Episode Title: State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI
- Show: Lex Fridman Podcast
- Guest 1 Role: Post-training lead at the Allen Institute for AI (Ai2)
- Guest 1 Book: The RLHF Book
- Guest 2 Book 1: Build a Large Language Model (From Scratch)
- Guest 2 Book 2: Build a Reasoning Model (From Scratch)
Nathan Lambert and Sebastian Raschka are machine learning researchers, engineers, and educators. Nathan is the post-training lead at the Allen Institute for AI (Ai2) and the author of The RLHF Book. Sebastian Raschka is the author of Build a Large Language Model (From Scratch) and Build a Reasoning Model (From Scratch).
Thank you for listening ❤ Check out our sponsors: https://lexfridman.com/sponsors/ep490-sc
See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc.
Transcript:
https://lexfridman.com/ai-sota-2026-transcript
CONTACT LEX:
Feedback – give feedback to Lex: https://lexfridman.com/survey
AMA – submit questions, videos or call-in: https://lexfridman.com/ama
Hiring – join our team: https://lexfridman.com/hiring
Other – other ways to get in touch: https://lexfridman.com/contact
SPONSORS:
To support this podcast, check out our sponsors & get discounts:
Box: Intelligent content management platform.
Go to https://box.com/ai
Quo: Phone system (calls, texts, contacts) for businesses.
Go to https://quo.com/lex
UPLIFT Desk: Standing desks and office ergonomics.
Go to https://upliftdesk.com/lex
Fin: AI agent for customer service.
Go to https://fin.ai/lex
Shopify: Sell stuff online.
Go to https://shopify.com/lex
CodeRabbit: AI-powered code reviews.
Go to https://coderabbit.ai/lex
LMNT: Zero-sugar electrolyte drink mix.
Go to https://drinkLMNT.com/lex
Perplexity: AI-powered answer engine.
Go to https://perplexity.ai/
OUTLINE:
(00:00) – Introduction
(01:39) – Sponsors, Comments, and Reflections
(16:29) – China vs US: Who wins the AI race?
(25:11) – ChatGPT vs Claude vs Gemini vs Grok: Who is winning?
(36:11) – Best AI for coding
(43:02) – Open Source vs Closed Source LLMs
(54:41) – Transformers: Evolution of LLMs since 2019
(1:02:38) – AI Scaling Laws: Are they dead or still holding?
(1:18:45) – How AI is trained: Pre-training, Mid-training, and Post-training
(1:51:51) – Post-training explained: Exciting new research directions in LLMs
(2:12:43) – Advice for beginners on how to get into AI development & research
(2:35:36) – Work culture in AI (72+ hour weeks)
(2:39:22) – Silicon Valley bubble
(2:43:19) – Text diffusion models and other new research directions
(2:49:01) – Tool use
(2:53:17) – Continual
- LLMs
- Nathan Lambert and Sebastian Raschka
- Large Language Models: AI systems trained on vast amounts of text data that can understand and generate human-like language.
- Scaling Laws
- Nathan Lambert and Sebastian Raschka
- Scaling Laws: Principles that describe how AI model performance improves predictably as you increase factors like data size or computing power.
- Agents
- Nathan Lambert and Sebastian Raschka
- Agents: AI systems designed to perform tasks autonomously, often by interacting with their environment or other software.
- AGI
- Nathan Lambert and Sebastian Raschka
- Artificial General Intelligence: A hypothetical type of AI that possesses human-like cognitive abilities across a wide range of tasks.
#500 – Khabib Nurmagomedov: Dagestan, MMA, UFC, Islam, Conor, Fedor & Football
WithKhabib Nurmagomedov
4 h 24#496 – FFmpeg: The Incredible Technology Behind Video on the Internet
WithJean-Baptiste Kempf, Kieran Kunhya
Same guests, other shows
1 h 19
1 h 08State of LLMs 2026: RLVR, GRPO, Inference Scaling — Sebastian Raschka
WithSebastian Raschka
1 h 28Open Source AI Strikes Back — Inside Ai2’s OLMo 3 ‘Thinking"
WithNathan Lambert, Luca Soldaini
Questions people ask
- Who are the guests on this episode?
- The guests are Nathan Lambert, post-training lead at the Allen Institute for AI, and Sebastian Raschka, author of several AI books.
- What topics will be covered?
- The episode will cover the state of AI in 2026, including LLMs, coding, scaling laws, China's AI, agents, GPUs, and AGI.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Lex Fridman.