Lex Fridman

AItechnologyscience

AI researcher at MIT and host of the Lex Fridman Podcast featuring long-form conversations about AI, science, technology, and the nature of intelligence.

Website @lexfridman

Latest Episodes

#491 – OpenClaw: The Viral AI Agent that Broke the Internet – Peter Steinberger

Peter Steinberger discusses the creation of OpenClaw, an open-source AI agent framework that became the fastest-growing project in GitHub history. The conversation covers how the project went viral, its self-modifying AI agent capabilities, security concerns, and the broader implications of AI agents for programming and software development. Steinberger also shares his life story, discusses acquisition offers from OpenAI and Meta, and offers his perspective on whether AI will replace programmers.

Self-modifying AI agents that can edit their own code represent a paradigm shift in how software is built and maintainedAI agents are predicted to replace approximately 80% of traditional apps by handling tasks directly rather than through conventional interfacesThe viral success of open-source AI projects depends on community engagement and solving real developer pain points at the right moment

Feb 12, 2026

#490 – State of AI in 2026: LLMs, Coding, Scaling Laws, China, Agents, GPUs, AGI

Lex Fridman hosts machine learning researchers Nathan Lambert and Sebastian Raschka for a comprehensive discussion on the current state of AI heading into 2026, covering topics from the US-China AI race and comparisons of leading LLMs to scaling laws, training methodologies, and the path toward AGI. The conversation dives deep into technical topics like pre-training, post-training with RLHF, open vs closed source models, and the evolving transformer architecture, while also addressing practical concerns like the best AI tools for coding and advice for beginners entering the field. They also discuss the intense work culture in Silicon Valley AI labs, emerging research directions like text diffusion models and tool use, and whether scaling laws are still holding or have plateaued.

AI scaling laws are not dead but are evolving, with post-training and inference-time compute becoming increasingly important alongside traditional pre-training scalingThe competition between ChatGPT, Claude, Gemini, and Grok is intensifying with each model excelling in different domains, particularly in coding where AI tools are rapidly transforming software developmentOpen source AI models are closing the gap with closed source alternatives, with the US-China AI race adding geopolitical complexity to questions of model access and development strategy

Feb 1, 2026

#489 – Paul Rosolie: Uncontacted Tribes in the Amazon Jungle

3:14:26

Paul Rosolie joins Lex Fridman to discuss his lifelong dedication to protecting the Amazon rainforest, including encounters with uncontacted tribes, the biodiversity and dangers of the jungle, and his conservation work through Junglekeepers. The conversation explores the fragility of the Amazon ecosystem, the ethics of engaging with isolated indigenous communities, and what modern civilization can learn from people who live in harmony with nature. Rosolie also discusses his new book 'Junglekeeper' and the urgent threats facing the world's largest tropical rainforest.

Uncontacted tribes in the Amazon represent some of the last truly isolated human societies and their protection is inseparable from rainforest conservationDirect immersion in the jungle reveals an interconnected web of life that cannot be fully understood or preserved from a distanceThe Amazon rainforest faces existential threats from deforestation and development, requiring grassroots conservation efforts like Junglekeepers to defend it

Jan 13, 2026

#488 – Infinity, Paradoxes that Broke Mathematics, Gödel Incompleteness & the Multiverse – Joel David Hamkins

Lex Fridman interviews mathematician and philosopher Joel David Hamkins about the nature of infinity, foundational paradoxes in mathematics, Gödel's incompleteness theorems, and the concept of a mathematical multiverse. The conversation spans topics from Russell's paradox and the halting problem to the Continuum Hypothesis, surreal numbers, and Conway's Game of Life. Hamkins, the top-rated user on MathOverflow, shares his perspective on the philosophical implications of set theory and the distinction between mathematical truth and proof.

Gödel's incompleteness theorems reveal a fundamental gap between truth and provability, showing that any sufficiently powerful mathematical system contains true statements it cannot proveThe Continuum Hypothesis and other undecidable problems suggest a mathematical multiverse where different set-theoretic universes can have different but equally valid truthsInfinity is not a single concept but a rich hierarchy of different sizes and structures, leading to deep paradoxes that reshaped the foundations of mathematics

Dec 31, 2025

#487 – Irving Finkel: Deciphering Secrets of Ancient Civilizations & Flood Myths

Irving Finkel, a curator at the British Museum and expert in ancient Mesopotamian languages, joins Lex Fridman to discuss the origins of human language, the development and decipherment of cuneiform writing, and his groundbreaking discovery of a tablet containing a flood narrative that predates the biblical Noah story. The conversation explores how ancient civilizations recorded their knowledge, the art of translating dead languages, and the significance of Mesopotamian mythology and its connections to later religious texts. Finkel also shares his controversial perspectives on sites like Göbekli Tepe and reflects on what ancient tablets reveal about the universality of human experience.

A Mesopotamian flood tablet predates the biblical Noah narrative and included technical instructions for building a circular arkCuneiform writing systems evolved from practical record-keeping into rich literary and mythological traditions across multiple ancient civilizationsDeciphering ancient languages reveals that fundamental human concerns and storytelling impulses have remained remarkably consistent across millennia

Dec 12, 2025

Get Lex Fridman's insights in your inbox

Subscribe and follow Lex Fridman to receive weekly AI-curated digests.