The Best AI Podcasts for Staying Informed

Podcasts are how I stay current with AI without burning my eyes out reading. I listen on walks, drives, and during workouts at 2x speed. Below are the shows I actually keep in my queue — broken down by what they're best for.

Definition

An AI podcast is an audio show — typically interview, panel, or solo monologue — covering artificial intelligence research, products, business, and culture, released on a regular schedule.

TL;DR

Latent Space and Cognitive Revolution are the best podcasts for AI builders and engineers.
Dwarkesh Patel runs the deepest, longest interviews with frontier researchers — required listening.
The AI Daily Brief is the best short-form daily podcast.
Lex Fridman, Hard Fork, and No Priors are the best for general AI culture and business.
Practical AI is the best for actually doing something with what you learn.

How I Picked These

I listen to roughly 30 podcasts on rotation. The ones below are the small set that pay rent in my queue. Three filters:

Substance — does the host actually understand the topic, or are they reading off a card?
Guest quality — are they pulling top-tier guests or recycling the LinkedIn-influencer circuit?
Editing — does the show waste my time with five minutes of intro, ads, and "what's up everybody"?

Anything that fails one of those gets cut. The list below survives.

Best Long-Form Interview Podcasts

These are the heavy hitters — 90 minutes to 4 hours, with researchers, founders, and practitioners.

1. Dwarkesh Podcast

Host: Dwarkesh Patel
Cadence: Weekly-ish (134+ episodes since 2020)
Best for: People who want the smartest, deepest AI interviews on the internet
Why it's worth it: Dwarkesh has become the default place where Demis Hassabis, Dario Amodei, Mark Zuckerberg, and Tony Blair go for long-form conversations about AI. His prep is insane — he reads everything the guest has written. Recent 2026 episodes include Dario Amodei (Anthropic), Dylan Patel (SemiAnalysis on the three bottlenecks to AI compute), and Michael Nielsen on scientific progress. If you only listen to one AI podcast, this is it.

2. Latent Space Podcast

Hosts: swyx (Shawn Wang) and Alessio Fanelli
Cadence: Weekly
Best for: AI engineers and people building products
Why it's worth it: The companion to the Latent Space newsletter. Every episode is a working interview with someone shipping something — eval frameworks, agent architectures, model post-training. The signal density is the highest of any builder-focused podcast.

3. No Priors

Hosts: Sarah Guo and Elad Gil
Cadence: Weekly
Best for: Founders, investors, and anyone tracking AI as a business
Why it's worth it: Sarah and Elad are both top-tier investors and they ask the questions other interviewers don't think of — about distribution, defensibility, and what actually makes an AI business work. Guests range from Sam Altman to lesser-known frontier-lab researchers.

4. Lex Fridman Podcast (AI episodes)

Host: Lex Fridman
Cadence: 2-4x per month
Best for: General audience, philosophical AI conversations
Why it's worth it: Skip the non-AI episodes. But when Lex has on someone like Demis Hassabis, Andrej Karpathy, or Sam Altman, it's a 3-4 hour deep dive that consistently produces the most-clipped quotes in the field. Editorially patchy, but the highs are very high.

Best Short-Form Daily Podcasts

When you want a 10-25 minute brief on what happened.

5. The AI Daily Brief

Host: Nathaniel "NLW" Whittemore
Cadence: Daily (Mon-Fri) — 984+ episodes as of May 2026
Best for: A daily AI news briefing for busy people
Why it's worth it: Nathaniel takes the day's biggest AI story and gives a smart, edited 20-minute take. He's especially good at corporate strategy and macro narratives. Running daily since 2023 with consistent quality. It's the AI equivalent of a smart morning radio show.

6. This Day in AI Podcast

Hosts: Michael and Chris Sharkey
Cadence: Weekly — 141+ episodes as of 2026
Best for: Hour-long news-and-vibes conversation between two builders
Why it's worth it: The Sharkey brothers cover every major model release in real time — Opus 4.6, Codex 5.3, GLM-5, Gemini drops, agentic workflows. Less rigorous than Latent Space, but a fun "two-friends-talking" format if you want news you can listen to like a conversation.

7. Marketplace Tech (AI segments)

Host: Lily Jamali
Cadence: Daily, AI segments roughly 2x per week
Best for: AI in the context of the broader economy
Why it's worth it: Marketplace's tech show consistently has sharp AI reporting that puts it in economic context — labor markets, productivity, regulation. Short, well-edited, NPR-quality.

Tip

For a daily AI listen, alternate The AI Daily Brief on weekdays with one long interview on Saturdays. That cadence gives you both breadth and depth without overload.

Best Podcasts for AI Engineers and Builders

If you ship AI products, these three are mandatory.

8. The Cognitive Revolution

Host: Nathan Labenz
Cadence: 2-3x per week
Best for: Practitioners, researchers, and people who want to understand model behavior
Why it's worth it: Nathan red-teams models for a living and his interviews go deep on capabilities, alignment, and what frontier models can actually do. His "tool use" episodes and his interviews with people like Riley Goodside are some of the best AI content anywhere.

9. Practical AI

Hosts: Daniel Whitenack and Chris Benson
Cadence: Weekly
Best for: ML engineers and applied AI practitioners
Why it's worth it: True to its name. Less hype, more "here's how we actually solved this engineering problem." If you're shipping ML in production, this is more useful than most of the marquee podcasts.

10. AI in Business

Host: Daniel Faggella (Emerj)
Cadence: Weekly — 1,057+ episodes (running since 2013)
Best for: Non-technical business leaders aligning AI with strategy and ROI
Why it's worth it: Faggella interviews Fortune 500 AI executives and unicorn AI founders on practical adoption. Less hype than most consumer podcasts, more "what actually moves the needle in enterprise AI." One of the longest-running AI podcasts on the internet.

11. The TWIML AI Podcast

Host: Sam Charrington
Cadence: 2x per week
Best for: Researchers and ML engineers
Why it's worth it: Sam has been doing this since 2016 — one of the longest-running ML podcasts. He's interviewed nearly every working ML researcher of note. The episodes lean technical and assume some baseline ML knowledge — exactly right for the audience.

12. Eye on AI

Host: Craig S. Smith (former NYT correspondent)
Cadence: Biweekly
Best for: Industry context with a journalist's framing
Why it's worth it: Craig's NYT background shows in his interview discipline. Recent 2026 episodes include Sriram Raghavan (IBM Research VP of AI), Steffen Cruz of Macrocosmos on training frontier models without data centers, and conversations on how human judgment shapes AI systems. Steady, smart, never sensational.

Best Podcasts for AI Business and Strategy

For people who care more about market impact than gradient updates.

13. Hard Fork

Hosts: Kevin Roose and Casey Newton
Cadence: Weekly
Best for: AI in culture, journalism, and the news cycle
Why it's worth it: Kevin and Casey are both excellent tech journalists. Their riffs on the week's news are smarter and funnier than most AI podcasts, and they get good guests. Best paired with The Information's reporting if you want to stay ahead of corporate moves.

14. Lenny's Podcast (AI episodes)

Host: Lenny Rachitsky
Cadence: Weekly — 320+ episodes
Best for: PMs, founders, and growth folks tracking how AI is reshaping product
Why it's worth it: Recent 2026 standouts include Elena Verna on Lovable hitting $200M ARR in a year, Alexander Embiricos on OpenAI Codex's product architecture, and Claire Vo's "How I AI" sister show on agent workflows. Lenny gets product-level conversations the AI-only podcasts don't.

15. The All-In Podcast (AI episodes)

Hosts: Chamath, Sacks, Friedberg, Calacanis
Cadence: Weekly
Best for: VC and political-economic angle on AI
Why it's worth it: Politics aside — and there's a lot of politics — when these four discuss AI investment, regulation, and corporate moves, it's some of the sharpest commentary you'll find. Skip the segments that aren't about AI.

16. Acquired (AI episodes)

Hosts: Ben Gilbert and David Rosenthal
Cadence: Bi-weekly, 3-5 hours each
Best for: Long-form business history of AI companies
Why it's worth it: Their episodes on NVIDIA, OpenAI, and the history of TSMC are the definitive long-form business histories. Marathon listens but absolutely worth the time investment.

Best Research and Frontier Podcasts

When you want to actually understand papers and capabilities.

17. Machine Learning Street Talk

Hosts: Tim Scarfe and rotating co-hosts
Cadence: Weekly-ish
Best for: Hard-technical AI listeners
Why it's worth it: The most rigorous AI podcast — episodes on AGI, neuroscience, cognitive architectures, and the philosophy of intelligence. Not for beginners. Pair with strong coffee.

18. Last Week in AI Podcast

Hosts: Andrey Kurenkov and Jeremie Harris
Cadence: Weekly
Best for: Comprehensive weekly news with technical context
Why it's worth it: Two hosts, two hours, every important AI story of the week — research, business, policy, hardware. The companion to the newsletter. If you only have time for one weekly catch-up, this is it.

19. 80,000 Hours Podcast (AI episodes)

Host: Rob Wiblin and others
Cadence: 2-4x per month
Best for: AI safety, alignment, and the long-term view
Why it's worth it: The best podcast for serious conversations about AI safety, governance, and what's at stake. The episodes with Ajeya Cotra, Holden Karnofsky, and Paul Christiano are some of the most thoughtful AI content anywhere.

Info

Most of these podcasts post their episodes to YouTube as well. If you absorb information better with visuals (slides, charts, code), watch instead of listen — and use 1.5x to 2x speed once you're warmed up to the host's cadence.

How I Listen

A few habits that make this volume sustainable:

2x speed for everyone except Lex (he's already slow at 1x)
Skip ads in Overcast or Pocket Casts — saves 10-15% per episode
Triage by guest — I don't listen to every episode, only the ones with guests I want to hear
Take notes — I keep a Notion database of useful clips and timestamps

Listening to AI podcasts is genuinely the highest-leverage way I learn about this field. You're getting 2-4 hour conversations with people who shape the industry, for free, on your morning walk.

Podcasts I've Quietly Removed

A few formerly-great AI podcasts have devolved into hype shows or pivoted to chasing crypto-style audience growth. The tell: when every episode is "this changes EVERYTHING" or "AI will replace your job by Tuesday." Trust your ear. If you're not learning, unsubscribe.

FAQ

What's the best AI podcast for absolute beginners?

The AI Daily Brief by Nathaniel Whittemore. It's short, well-edited, and assumes no technical background. After a few weeks of that, add Hard Fork for context and culture. Once you're comfortable, graduate to Dwarkesh and No Priors.

What's the best AI podcast for engineers and builders?

Latent Space, full stop. Then The Cognitive Revolution and Practical AI. These three cover almost the entire AI engineering surface area — agents, evals, fine-tuning, deployment.

How can I keep up with AI podcasts without spending hours per day listening?

Three rules. Listen at 2x speed. Triage by guest, not by show — only listen if the guest is someone you want to hear. And do most listening during commutes, walks, and workouts so it doesn't compete with your work hours.

Are there any good AI podcasts hosted by women?

Yes. No Priors with Sarah Guo is one of the best podcasts on this list. Marketplace Tech is hosted by Lily Jamali. Plus, many of the deepest AI episodes feature researchers like Helen Toner, Ajeya Cotra, and Sarah Bird as guests across most of these shows.

Should I listen to podcasts or read newsletters to stay informed?

Both, for different reasons. Newsletters are faster and more efficient for news. Podcasts give you context, opinion, and depth that text can't match. The combo of one daily newsletter plus one or two weekly podcasts is the sweet spot for most people.

The list above gives you a complete AI podcast diet — daily news, builder depth, business strategy, research frontier. Pick three, give them a month, and trust your queue. The goal is to feel current without feeling overwhelmed.