20 AI Product Podcasts Worth Your Time If You're Scaling LLM Capabilities

There are hundreds of AI podcasts. Most cover announcements. Very few cover the work — the measurement problems, the adoption failures, the architectural decisions, and the uncomfortable strategic questions that determine whether an LLM-powered product survives its first year.

This list is organized by the challenge you're facing. If you're trying to figure out why your AI product has usage but no retention, there's a podcast for that. If you need to understand why your agent deployment is failing silently, there's one for that too. Two standout episodes per show so you can test before you commit.

---

For measuring AI product impact

1. Product Impact Podcast

Best for: Teams that need frameworks for proving AI creates value — not just activity. We cover measurement, adoption, defensibility, and the behavioral layer where enterprise AI value is won or lost.

Guests: Robert Brunner (Apple, Ammunition), Helen Edwards (Artificiality Institute), Juan Sequeda (ServiceNow), Teresa Torres (Product Talk), Devi Parikh (Yutori).

Standout episodes: S02E01 — Why Your AI Metrics Are Lying to You (the Bullseye framework) · S02E05 — The Human Impact of AI We Need to Measure (cognitive sovereignty with Helen Edwards)

Listen: Spotify · Apple · YouTube

2. Lenny's Podcast

Best for: Product managers connecting AI capabilities to growth, retention, and monetization. Lenny Rachitsky consistently gets the most senior product leaders in tech to explain how they actually make decisions — not how they talk about decisions on stage.

Guests: CPOs and VPs of Product from Figma, Notion, Duolingo, Airbnb, Stripe, and the companies defining AI-native product management.

Standout episodes: His deep-dive on AI product metrics with the Amplitude team · His conversation with the Notion AI team on how they decide what to automate vs. what to assist

Listen: Spotify · Apple

---

For AI product strategy and venture perspective

3. AI + a16z

Best for: Understanding where billions in AI investment thesis are pointing. The Andreessen Horowitz perspective on infrastructure, developer tooling, and the application layer — reflecting real capital allocation, not speculation.

Standout episodes: Their analysis of AI application layer economics and why most wrappers die · Their framework for evaluating AI startup defensibility beyond model capability

Listen: Spotify · Apple

4. No Priors

Best for: AI startup founders who need funding, competitive dynamics, and go-to-market to be real, not theoretical. Sarah Guo and Elad Gil bring on founders building actual businesses.

Standout episodes: Their breakdown of why enterprise AI sales cycles are lengthening · Their episode on the economics of fine-tuning vs. prompting at scale

Listen: Spotify · Apple

5. The a16z Podcast (Lightcone)

Best for: The venture lens on which AI business models survive the shakeout. Distinct from AI + a16z — Lightcone covers broader tech with AI as the dominant reshaping force.

Standout episodes: Why AI startup unit economics are structurally different from SaaS · Their analysis of the infrastructure vs. application layer capital allocation imbalance

Listen: Spotify · Apple

---

For AI-native business building

6. The Startup Ideas Podcast (Greg Isenberg)

Best for: Founders who want concrete AI startup ideas, distribution strategies, and the modern playbook for building AI-native businesses. Greg Isenberg (CEO of Late Checkout, former advisor to Reddit and TikTok) publishes twice weekly with specific, actionable business ideas — not abstract trend analysis. He consistently finds the intersection of AI capability and distribution advantage that most strategy podcasts miss.

Standout episodes: His 30-step playbook for building a modern SaaS company using AI agents and media · His episode on why building an MCP server in 2026 is like building for mobile in 2010 — early movers will own AI-native distribution channels

Listen: Spotify · Apple · YouTube

7. AI in Business

Best for: Enterprise leaders connecting AI to business outcomes, not capability demonstrations. Dan Faggella (Emerj) zeroes in on where AI actually creates revenue and reduces cost — with the receipts.

Standout episodes: Why enterprise AI pilots fail to scale — directly relevant to the Proof Gap pattern · His episode on AI procurement: how enterprise buyers actually evaluate AI vendors

Listen: Spotify · Apple Podcasts

---

For AI engineering and deployment

8. Latent Space

Best for: Engineering leaders making RAG vs. fine-tuning vs. prompt engineering decisions. Swyx and Alessio Fanelli cover problems practitioners face daily — evaluation frameworks, inference optimization, and the infrastructure that works in production vs. the infrastructure that works in demos.

Standout episodes: Why RAG architectures fail in production and what top teams do differently · Their episode on building evaluation pipelines that catch regressions before users do

Listen: Spotify · Apple

9. Gradient Dissent (Weights & Biases)

Best for: ML engineering leaders who need to hear how teams at DeepMind, Meta AI, and Anthropic actually build and ship. Lukas Biewald runs Weights & Biases — used by virtually every ML team — which gives him a front-row seat and unusually specific questions about what breaks at scale.

Standout episodes: Their episode on inference cost optimization when your bill exceeds your revenue · The conversation on data quality pipelines — why garbage in still produces confident garbage out

Listen: Spotify · Apple

10. Practical AI

Best for: Teams moving from POC to production. Chris Benson and Daniel Whitenack focus on the engineering decisions that determine whether an AI system ships or stays in a notebook forever.

Standout episodes: LLM observability in production — monitoring what your model does when nobody watches · Their breakdown of when to fine-tune, when to RAG, and when to just write a better prompt

Listen: Spotify · Apple

---

For the big picture and uncomfortable questions

11. Dwarkesh Patel

Best for: The 60-120 minute conversation with a frontier AI leader that nobody else gets. Dwarkesh asks sharper, more confrontational questions than any other interviewer in AI — and his guests can't hide behind talking points because the format is too long for that to work.

Standout episodes: His 2026 Dario Amodei interview on LLM coding ability and continual learning · His conversation with Jeff Dean on what Google got wrong about transformer scaling

Listen: Spotify · YouTube

12. The Cognitive Revolution

Best for: The middle ground between technical depth and strategic breadth that most podcasts can't hold. Nathan Labenz and Erik Torenberg interview builders, researchers, and investors across the full stack.

Standout episodes: Their episode on AI evaluation frameworks for non-ML-engineers · Their conversation on why AI research benchmarks and production performance are diverging

Listen: Spotify · Apple

13. Hard Fork (NYT)

Best for: Understanding how AI is perceived outside the tech bubble. Kevin Roose and Casey Newton cover AI from the perspective of impact on real people — the audience your product actually serves, not the engineers who built it.

Standout episodes: Their coverage of the AI workforce displacement debate · Their episode on why consumer AI trust is declining even as capability increases

Listen: Spotify · Apple

14. Lex Fridman Podcast

Best for: Long-form philosophical conversations that reframe how you think about AI at the 10-year horizon — not what you ship this quarter, but what the technology means for the species.

Standout episodes: Any Ilya Sutskever conversation — the most consequential AI thinker who rarely speaks publicly · His Sam Altman interview on the gap between OpenAI's public narrative and internal reality

Listen: Spotify · YouTube

---

For enterprise adoption and governance

15. AI Product Leader

Best for: Product managers transitioning into AI-native roles. Polly Allen interviews AI product leaders about career paths, team structure, and the skills that separate AI PMs from traditional PMs.

Standout episodes: Building AI product teams from scratch — hiring, structure, and evaluation · Her episode on when to build AI in-house vs. when to buy

Listen: Spotify

16. The AI Agents Podcast

Best for: Teams deploying AI agents in production who need both the technical reality and the economic honesty. Covers the graduated autonomy pattern and why most agent deployments fail within 90 days.

Standout episodes: Why 90% of legacy agents fail within weeks — what the surviving 10% do differently · Their breakdown of agent observability: what to monitor when the agent acts without you watching

Listen: Spotify

17. Eye on AI

Best for: The regulatory and governance perspective most product podcasts skip entirely. Craig Smith covers AI policy, safety, and the regulatory landscape that determines what you're allowed to ship — and when that changes by jurisdiction.

Standout episodes: EU AI Act's impact on enterprise deployment — the compliance reality most teams haven't planned for · His episode on AI liability: who's responsible when the agent causes harm

Listen: Spotify · Apple

---

For AI research translated to practice

18. TWIML AI (This Week in Machine Learning)

Best for: Translating ML research into product decisions. Sam Charrington bridges the gap between what researchers publish and what practitioners need to ship — without dumbing it down or losing the nuance that matters for implementation.

Standout episodes: Why academic benchmarks mislead product teams · The conversation on production ML monitoring that caught problems academic evals missed entirely

Listen: Spotify · Apple

19. Last Week in AI

Best for: Staying current without drowning. A weekly digest that covers the most important AI developments — useful when you need to know what happened but don't have time to follow 50 sources. The signal-to-noise ratio is the highest of any news-format AI podcast.

Standout episodes: Their coverage of the Q1 2026 agentic AI failure cluster · Their analysis of why AI model costs are declining faster than anyone predicted and what it means for product pricing

Listen: Spotify · Apple

20. Machine Learning Street Talk

Best for: The technical depth other podcasts avoid. Tim Scarfe and guests go deep on architecture decisions, training methodology, and theoretical foundations. Not for casual listening — for when you need to understand why at the engineering level, not just what shipped.

Standout episodes: Their dissection of why scaling laws are slowing and what it means for product roadmaps · The conversation on evaluation methodology — why most AI benchmarks are measuring the wrong thing

Listen: YouTube · Spotify

---

The uncomfortable truth about AI podcasts: most are optimized for downloads, not for your product outcomes. The 20 above are the ones where I consistently learn something that changes what I build or how I advise teams at PH1 Research. If you're scaling LLM capabilities and need signal over noise, start with the category that matches your hardest problem right now — and listen to one episode before subscribing to all of them.

---

Related:
- How to Measure AI Product Impact: The Bullseye Framework
- Why AI Capability Is No Longer Defensible
- The Agentic Era: What AI Agents Are and How They Change Work
- Enterprise Context Is the AI Moat Nobody Built