YouTube Channels That Explain AI Properly (2026): 15 That Teach Mechanism, Not Hype

Doom Debates: AI debates that must be resolved before the world ends

doomdebates.com·Liron Shapira · Article

A Substack newsletter by Liron Shapira focused on critical debates surrounding artificial intelligence and existential risks. The publication addresses fundamental questions about AI development and safety that the author argues must be resolved before catastrophic outcomes occur, attracting over 1,000 subscribers interested in AI policy and philosophical issues at the intersection of technology and existential risk.

Nopriors

nopriors.com·Article

Nopriors is a newsletter or content platform focused on artificial intelligence and technology topics. The site serves as a hub for curated discussion and analysis of AI developments, trends, and implications.

Stratechery by Ben Thompson

stratechery.com·Article

Stratechery is Ben Thompson's newsletter and platform analyzing business strategy and the impact of technology. This week's content covers tech earnings from Meta, Microsoft, Amazon, and Google, examining their divergent approaches to AI infrastructure investment and the market's contrasting reactions. Notable coverage includes OpenAI's response to Apple's trade secret lawsuit regarding hardware division employees, analysis of frontier AI model costs and monetization potential, and discussion of open-weights AI models from China and their implications for industry structure. The site also features companion podcasts including Dithering with John Gruber, Sharp Tech on platform strategy, and Greatest of All Talk on sports, maintaining a focus on how technological and business decisions reshape markets and industries.

ChinaTalk: Deep Coverage of Technology, China, US Policy, and War

chinatalk.media·Jordan Schneider · Article

ChinaTalk is a Substack newsletter with over 77,000 subscribers run by Jordan Schneider, offering original analysis and interviews with leading thinkers and policymakers on technology, China, US policy, and geopolitical conflict. The publication is cited by prominent analysts and researchers including Noah Smith, Jon Y (Asianometry), and Dylan Patel (SemiAnalysis) for providing fair, well-researched, and balanced coverage of China-related topics.

The TWIML AI Podcast with Sam Charrington

twimlai.com·@samcharrington · Article

A long-running podcast series hosted by Sam Charrington featuring expert conversations on machine learning, artificial intelligence research, tools, and real-world deployment strategies. The show covers emerging AI trends, research breakthroughs, and practical applications through interviews with researchers, engineers, and practitioners in the field. With over 770 episodes, it serves as a comprehensive resource for staying current with developments in AI and machine learning.

Latent Space

latent.space·Latent.Space · Article

A podcast and newsletter covering AI engineering news, research papers, and interviews with founders and builders at the cutting edge of foundation models, code generation, multimodality, AI agents, and GPU infrastructure. Episodes feature exclusive interviews from leaders at OpenAI, Anthropic, Meta, Databricks, Modular, and other major labs, with show notes available at latent.space. The publication reaches over 10 million readers and listeners annually and aims to provide both breaking news on current developments and introductions to technologies entering mainstream use within three months.

80,000 Hours Podcast

80000hours.org·Article

A podcast featuring in-depth conversations on the world's most pressing problems and strategies to solve them. Hosted by Robert Wiblin, Luisa Rodriguez, Zershaaneh Qureshi, and Tom Reed, the show includes a curated crash course on transformative AI covering both potential benefits and risks. The hosts bring backgrounds in economics, international development, mathematics, philosophy, AI safety research, and policy, drawing on their work at organizations including 80,000 Hours, the Centre for Effective Altruism, and the UK's AI Security Institute.

The Cognitive Revolution

cognitiverevolution.ai·Article

A weekly podcast hosted by Nathan Labenz featuring interviews with AI innovators and thought leaders exploring the near-term transformative impacts of artificial intelligence.

Google DeepMind

youtube.com·Google DeepMind · Channel

Google DeepMind is a research organization combining scientists, engineers, ethicists and others focused on developing AI systems safely and responsibly. The team works on foundational scientific and engineering challenges with the goal of creating breakthrough technologies that could advance science, transform work, and improve lives at scale. Their mission emphasizes responsible development toward artificial general intelligence (AGI).

Interconnects AI

interconnects.ai·Nathan Lambert · Article

A Substack newsletter by Nathan Lambert covering cutting-edge AI developments from inside frontier labs, balancing high-level concepts with technical depth. The publication attracts over 77,000 subscribers including leading engineers, researchers, and investors seeking insights grounded in practical experience rather than hype. Lambert's commentary is recognized for its accessibility and authority, drawing on his direct experience training AI models to provide informed perspectives on open-source AI and the current state of the field.

AXRP - the AI X-risk Research Podcast

axrp.net·Article

AXRP is a podcast hosted by Daniel Filan featuring conversations with researchers working on artificial intelligence existential risk reduction. The show explores why researchers conduct their work and how it might help prevent AI from causing an existential catastrophe that could permanently curtail humanity's future potential. Episodes cover topics including mechanistic interpretability, alignment, AI governance, evaluations, and forecasting. Full episode transcripts are available on the website, and the podcast can be accessed through standard podcast providers or via RSS feed.

Machine Learning Street Talk (MLST)

mlst.ai·Machine Learning Street Talk · Article

Machine Learning Street Talk is a media production company creating video, audio, and written content focused on machine learning, cognitive science, and cognitive philosophy. They operate a podcast that ranks as the highest-rated technical AI podcast on Spotify, and publish a Substack newsletter with over 2,000 subscribers.

The path, if starting from zero: nine hours plus one weekend

Most lists dump 20 channels and leave you to it. The sequence matters more than the set. 1. Get the shape. 3Blue1Brown, chapters 1 to 7. About 3 hours. 2. Build one yourself. Karpathy, "Let's build GPT from scratch". About 2 hours, plus a weekend actually typing it. 3. Build a real one. Umar Jamil, Transformer from scratch in PyTorch. About 3 hours. 4. Learn to judge claims. AI Explained, any three recent model-analysis videos. About 1 hour. 5. Go deep on one thing. GPU MODE or Stanford CS25, pick by interest. Ongoing. Steps 1 to 4 are roughly nine hours plus one weekend of typing. That puts you ahead of most people who have been reading AI Twitter daily for a year, which is the actual claim this list makes.

Dwarkesh Patel

youtube.com·Dwarkesh Patel · Channel

Dwarkesh Patel runs a YouTube channel featuring deeply researched, long-form interviews with thinkers, researchers, and innovators across fields including AI, philosophy, economics, and technology. The channel is known for in-depth conversations that explore ideas thoroughly, with guests ranging from AI researchers to economists to creative professionals.

Robert Miles AI Safety

youtube.com·Robert Miles AI Safety · Channel

A channel focused on artificial intelligence safety research and the challenges posed by advancing AI systems. Robert Miles covers how AI could go wrong and explores humanity's efforts to foresee problems and ensure safe, beneficial AI development, explained for a general audience rather than specialists.

Yannic Kilcher

youtube.com·Yannic Kilcher · Channel

Yannic Kilcher is a YouTuber who creates videos analyzing machine learning research papers, exploring programming topics, and discussing issues within the AI community and AI's broader societal impact. His channel focuses on making deep learning and AI research accessible to technical audiences while examining the ethical and policy implications of artificial intelligence development and deployment.

Machine Learning Street Talk

youtube.com·Machine Learning Street Talk · Channel

MLST is a technical AI podcast featuring interviews with leading AI researchers and experts. The show covers advanced AI research, alternative paths to AGI beyond large language models, cognitive science, computer science, neuroscience, mathematics, and philosophy of mind and language. The channel emphasizes diversity of ideas and rigorous exploration across multiple disciplines relevant to artificial intelligence development.

AI Explained

youtube.com·AI Explained · Channel

Channel covering major developments in artificial intelligence, particularly focused on the emergence of advanced AI systems approaching human-level intelligence. The creator authored Simple Bench, a tool for measuring and exposing gaps in reasoning capabilities between large language models and humans. Also developed LM Council, an AI-related project, and maintains a Patreon community called AI Insiders with over 1,000 members interested in AI topics and professional development in the field.

Alexander Amini

youtube.com·Alexander Amini · Channel

YouTube channel for Alexander Amini, a developer and researcher focused on artificial intelligence and machine learning topics. The channel features educational content and discussions related to AI technologies and learning resources.

Stanford Online

youtube.com·Stanford Online · Channel

Stanford Online is the university's portal for academic and professional education, operated by the Stanford Engineering Center for Global & Online Education. The platform offers degree programs, credit-bearing courses, professional certificates, and free open educational content created by Stanford faculty. It provides access to Stanford teaching and research globally, helping learners expand knowledge, advance careers, and enhance skills across multiple disciplines and learning levels.

GPU MODE

youtube.com·GPU MODE · Channel

GPU MODE is a reading group and community focused on GPU computing and artificial intelligence topics, created by Mark Saroufim and Andreas Köpf. The channel serves as a hub for supplementary content and discussion, with an active Discord community and a GitHub repository for shared resources and materials.

Umar Jamil

youtube.com·Umar Jamil · Channel

Channel by a machine learning engineer based in Milan, Italy, focused on teaching complex deep learning and machine learning concepts. The creator speaks multiple languages including English and Chinese.

Andrej Karpathy

youtube.com·Andrej Karpathy · Channel

YouTube channel for Andrej Karpathy, a prominent AI researcher and engineer known for his work in deep learning and computer vision. The channel features educational content and videos related to artificial intelligence and machine learning topics.

Mutual Information

youtube.com·Mutual Information · Channel

A YouTube channel producing educational videos on machine learning and related technical topics.

Artem Kirsanov

youtube.com·Artem Kirsanov · Channel

A YouTube channel by a Harvard neuroscience researcher exploring how neural systems learn and compute at the intersection of neuroscience, computer science, and mathematics. The channel focuses on understanding the fundamental mechanisms behind neural computation and learning across biological and artificial systems.

StatQuest with Josh Starmer

youtube.com·StatQuest with Josh Starmer · Channel

A YouTube channel dedicated to breaking down statistics, machine learning, data science, and AI into simple, understandable components. Rather than oversimplifying, the creator builds foundational understanding step-by-step so viewers develop genuine competence in these typically intimidating subjects. Each technique is presented as a combination of small, digestible steps that together form complex methodologies.

Welch Labs

youtube.com·Welch Labs · Channel

Welch Labs is a YouTube channel that creates animated educational content about artificial intelligence and machine learning concepts. The channel is known for making complex AI topics accessible through high-quality illustrations and visual explanations. The creator has recently released "The Welch Labs Illustrated Guide to AI," available for pre-order, which extends the channel's educational mission into book format.

3Blue1Brown - Math and Physics Animation Videos

youtube.com·3Blue1Brown · Channel

Grant Sanderson's channel features animated explanations of mathematics, physics, and computer science concepts, prioritizing visual intuition and clear perspective over rote memorization. Each video uses custom animation to break down complex or abstract ideas, making difficult problems accessible through interactive demonstrations and conceptual reframing. The channel emphasizes showing why mathematical and scientific concepts matter, not just how to compute them.