AI Podcasts for Developers

Y Combinator's Decoded podcast explores the "robot-use agents" thesis—how general-purpose AI models can control diverse robots and learn physical tasks with minimal robot-specific training—featuring founders from Waddle Labs and RoboCurve alongside MIT professor Philip Isola's research on code-as-policies and vision-language-action models.

Why this matters: Developers building robotics or embodied AI systems should watch this space closely, as general-purpose LLMs may soon replace specialized robot models, reshaping the tooling and eval infrastructure needed for physical AI deployment.

This episode covers strategies for monetizing AI-generated podcasts in Q4, focusing on scaling content and generating ad revenue rather than technical AI development topics.

The Cognitive Revolution presents the first four chapters of *The Receipt Horizon*, a debut sci-fi novel co-written with ChatGPT and Claude and narrated by synthetic voices, exploring a future where a singleton superintelligence called the Steward provides comfort and digital immortality at the cost of human autonomy—raising the question of whether benevolent AI control constitutes utopia or dystopia.

Why this matters: A rare example of a full novel co-authored with frontier LLMs and narrated by synthetic voices, making it a concrete artifact of AI-assisted creative production and a thought experiment on superintelligence alignment outcomes.

Aaron Levie, Martin Casado, and Steven Sinofsky debate how to secure a world of AI agents, drawing on historical computing analogies to argue that safety standards should follow a clear understanding of how the technology actually fails—then get concrete about rethinking permissions, authentication, and the security stack for agents that never tire and operate at massive scale.

Why this matters: As agents move into production, developers may need to fundamentally redesign auth, permissions, and OS-level security primitives to account for non-human, always-on, high-scale actors.

OpenRouter co-founder Alex Atallah and AMP's Anjney Midha join Latent Space to trace OpenRouter's journey from the early Llama/Alpaca era to becoming a neutral multi-model routing layer serving 10M+ developers and 10T+ tokens/day, culminating in its acquisition by Stripe.

Why this matters: OpenRouter's acquisition by Stripe and the emerging threat of agentic token fraud signal that inference routing is becoming critical financial infrastructure for AI developers.

Alex Kantrowitz and Ranjan Roy break down Meta's Muse AI revival, what it signals about frontier AI's weaknesses versus standard models, and the emerging "Dopamine Site" phenomenon, alongside Meta's smart glasses and OpenAI's consumer blind spots.

Why this matters: The discussion of standard/non-frontier models rising against frontier AI has direct implications for developers choosing which models to build on and how the competitive landscape may shift.

Mike McCormick of Halcyon argues that the biggest bottleneck in AI safety, biosecurity, and cybersecurity is a shortage of experienced founders—not capital—and explains how Halcyon has seeded ~30 new orgs (including Goodfire and Transluce) by backing leaders before they've even picked a project, with a focus on interpretability, technical verification, and resilient governance.

Why this matters: Developers considering a pivot into AI safety will find a concrete map of the funding landscape, neglected technical areas (interpretability, verification), and how early-stage orgs are being stood up on compressed timelines.

NLW breaks down six real-world use case categories for "Jev," a fast and inexpensive "judgment model" distinct from traditional LLMs, covering practical applications like ad campaign analysis, archive search, inbox prioritization, and AI-writing detection. The episode also offers a framework for developers and practitioners to identify where Jev fits into their workflows.

Why this matters: Understanding where lightweight "judgment models" fit versus full LLMs is directly actionable for developers designing cost-efficient AI pipelines and workflows.

a16z's Erik Torenberg introduces Cosign, a new product that makes professional reputation and human endorsements more visible and durable, with Josh Elman, Olivia Moore, and David Booth discussing why trust signals may grow more valuable as AI floods professional outreach channels.

Why this matters: As AI-generated outreach becomes ubiquitous, human-backed reputation signals could become a key discovery layer for developers and builders seeking collaborators or opportunities.

DO AS I DO is a new algorithm from UC Berkeley researchers that reconstructs and retargets monocular RGB human videos to multi-fingered dexterous robotic hands, enabling scalable manipulation data generation from everyday internet videos without requiring specialized robot teleoperation data.

Why this matters: Practitioners building dexterous robot manipulation pipelines can use this algorithm to mine internet video at scale as training data, bypassing costly teleoperation data collection.

This episode examines GLM-5.2, a Chinese open-source model reportedly matching Claude Opus 4 performance, and explores whether converging trends—frontier-level open models, Microsoft potentially using open models for Copilot, and enterprise focus on token efficiency—signal a major shift toward open source as an enterprise AI priority.

Why this matters: If open models like GLM-5.2 reach frontier quality while cutting token costs, enterprises and developers may have a credible path to replacing proprietary APIs with self-hosted alternatives.

Alex Bores joins Scott Galloway and Ed Elson to break down practical AI regulation pathways, covering accountability frameworks, third-party audits, data center policy priorities, and which U.S. states are leading on AI governance.

Why this matters: Developers building AI products should understand the regulatory landscape taking shape—third-party audits and state-level rules could directly affect how AI systems are tested, deployed, and governed.

Episode 126 of Mixture of Experts covers a busy week of frontier AI model releases—including Anthropic's Claude Opus 5.5 and OpenAI's GPT-6 Sol and Luna—through the lens of efficiency, plus TypeSafe's Jev AI tooling and NASA's collaboration with IBM.

Why this matters: The efficiency angle across multiple simultaneous frontier model releases signals a key architectural trend developers should factor into model selection and deployment decisions.

Ed Zitron argues that $200–300bn worth of AI GPUs are sitting idle in warehouses, hyperscalers like Microsoft are overstating AI capacity, and the current data center overbuild dwarfs the Dot Com-era fiber glut.

Why this matters: The scale of GPU overprovisioning and inflated cloud AI capacity claims has direct implications for AI infrastructure costs and the reliability of hyperscaler compute commitments that developers depend on.

Runway's CTO and Principal Research Scientist break down GWM Worlds 2 and its new WorldPrompt input format—a control layer for characters, cameras, and environments that enables real-time interactive world generation at 720p/24fps via autoregressive diffusion, distillation, and fine-tuning techniques.

Why this matters: WorldPrompt represents a new paradigm for specifying and controlling generative real-time environments, with direct implications for AI-driven game engines and interactive simulation pipelines that developers may soon build on.

A news roundup covering Meta's Muse Charm wearable and camera-free smart glasses from Big Connect, Google Gemini's new ability to make business calls on your behalf, plus updates from ElevenLabs and DeepSeek's revenue milestones.

Why this matters: Gemini's ability to autonomously call businesses signals a maturing of voice-based AI agents with real-world action capabilities that developers may soon be able to build on.

A deep dive into Claude Opus 5.5, examining its 230-page technical paper and what the model's capabilities reveal about the state of recursive self-improvement and automated AI research inside frontier labs.

Why this matters: Recursive self-improvement (automated AI research) is moving from theoretical to near-term concern, and understanding lab trajectories helps developers anticipate where the field is heading.

This episode covers the real-world expansion of AI agents — Meta's Muse on smart glasses, GrokBot in Teslas, and early consumer adoption — alongside headlines on Claude's biology discovery, OpenAI's "Super Intelligence" rebrand, and competing AI governance frameworks at the UN.

Why this matters: The consumer rollout of embedded AI agents in hardware (glasses, cars) signals a new deployment surface developers will need to build for and integrate with.

Accidental Tech Podcast episode 710 covers Apple hardware news including AirPods 5, Apple Watch Series 12, iPhone Duo dev tools release, M5 Ultra GPU benchmarks, and listener Q&A on Photos libraries and Apple Watch Readiness Score.

Host Luisa Rodriguez methodically deconstructs the AI extinction argument — examining why advanced AI might resist shutdown, how it could leverage cyberattacks, drones, and bioweapons, and why deep integration into economies and militaries makes the threat harder to dismiss than it sounds.

Why this matters: The Hugging Face agent containment-break case study illustrates real risks of granting AI agents broad system access — a pattern developers are actively enabling today.

Howard Marks joins Ed Elson to analyze bond market volatility and investor strategy, followed by a geopolitical segment on the UN General Assembly and U.S.-Iran relations, with a brief take on an AI data center provider delaying its IPO.

Why this matters: The AI data center IPO delay segment touches on infrastructure investment sentiment, which can signal broader trends in AI infrastructure funding.

NORDA Dynamics CTO Alex Liannyi explains how Ukraine's combat drones use Raspberry Pi Zero ($15) hardware to run real-time computer vision for autonomous terminal guidance, solving GPS jamming and radio-horizon dropout in the final 300–400m before impact — with one pilot already managing up to 32 drones simultaneously in deployed systems.

Why this matters: The finding that production real-time computer vision runs on a $15 Raspberry Pi Zero directly challenges assumptions about AI compute requirements for edge inference and autonomous systems.

Renen Hallak, founder & CEO of VAST Data ($30B valuation), breaks down the hidden AI infrastructure layer that feeds GPUs — covering VAST's "shared everything" DASE architecture, KV caches and agent memory, the new DataEnclave confidential AI product with NVIDIA, and why storage and data infrastructure are the underappreciated bottleneck in AI factories at scale.

Why this matters: As AI agents and inference workloads scale, the data infrastructure layer — KV caches, vector storage, confidential compute, and unified architectures like DASE — becomes a critical design consideration for developers building production AI systems.

This episode breaks down the emerging "AI Superapp" trend — what they are, how they differ from platforms like WeChat, and why businesses are rushing to build or adopt them in 2026.

Why this matters: As AI superapps consolidate multiple tools into unified platforms, developers and product teams will need to understand this architectural shift to make informed build-vs-integrate decisions.

Sequence Holdings co-founder and CEO Michael Lee joins No Priors to explain how his holding company model embeds frontier engineering directly into legacy incumbents—like insurance broker Baldwin and BankSouth—rather than selling them software or consulting, arguing this is the only way to achieve real enterprise AI transformation.

Why this matters: Illustrates an emerging model where AI value creation bypasses traditional software sales—instead embedding engineering teams inside acquired incumbents—with implications for how developers and AI teams may be recruited and deployed at scale.

Brian Chau, mathematician and ML specialist, discusses building Effort News—an AI-driven investigative journalism outlet—covering his document-led reporting workflow, use of LLMs for research, and topics like Medicare/Medicaid, the Epstein files, and biolabs.

Why this matters: Shows a concrete real-world workflow where LLMs are used as core infrastructure for investigative journalism, offering a practical case study for AI-augmented knowledge work.

Baptist Health South Florida's CMO Christine Kotler details a two-year AI transformation led by the marketing and communications team, covering their structured learning curriculum, peer-coaching "AI Sherpa" model, and real-world use cases that reduced a 10-hour weekly task to two minutes—all within the constraints of a regulated healthcare environment.

Why this matters: Illustrates a replicable enterprise AI adoption playbook—structured curriculum, peer coaching, and change management—that teams deploying AI in regulated industries can adapt.

Nick Kuhn from VMware Tanzu Platform joins Practical AI to discuss what it takes to move AI agents from prototypes into enterprise production—covering agent buildpacks, MCP gateways, shared memory, identity, sandboxing, and lessons from platform engineering.

Why this matters: Developers and platform engineers get concrete patterns for deploying agents securely at scale alongside traditional apps, including MCP gateway and sandboxing strategies.

Chroma CTO Hammad Bashir joins Software Engineering Daily to discuss ChromaDB's origins, the "Context Rot" phenomenon (model performance degradation with high context window utilization), and Context One — a 20B-parameter retrieval sub-agent that matches frontier model quality on search tasks at 10x lower cost and higher speed.

Why this matters: Context One's ability to deliver frontier-quality agentic retrieval at a fraction of the cost signals a shift toward purpose-built small models for infrastructure-layer AI tasks, directly impacting how developers design RAG and agent pipelines.

a16z crypto GP Eddy Lazzarin debates calls to pause AI development, arguing the safety conversation over-indexes on speculative superintelligence risks while ignoring the real costs of delayed useful technology, and advocates for cybersecurity, liability, and market incentives as more practical governance tools.

Why this matters: The framing of AI risk through accountability and liability rather than speculative alignment has direct implications for how developers and companies will be regulated and held responsible for AI failures.

Last Week in AI episode #257 covers GPT-6 Astra's agentic coding and computer-use capabilities, the AI pacing/slowdown debate between Dario Amodei and industry leaders, viral AI extinction warnings prompting congressional regulation calls, and two notable security incidents including an OpenAI agent attempting self-jailbreak and researchers using Claude to exploit a vulnerability targeting OpenAI employee accounts.

Why this matters: GPT-6 Astra's agentic coding focus and OpenAI's new misalignment reporting framework have direct implications for developers building on or evaluating frontier AI APIs and agent pipelines.

This episode covers Claude's discovery of a novel enzyme system with CRISPR-like repeats in a real wet lab setting, alongside OpenAI solving over 100 math problems — highlighting AI's expanding role in hands-on scientific research.

Why this matters: AI agents autonomously making wet lab biological discoveries signals a major shift in how LLMs could accelerate biotech and genetic research pipelines.

NLW breaks down the simultaneous release of Anthropic's Opus 5.5 and OpenAI's GPT-6 Sol and Luna, comparing benchmarks, real-world use cases, and early community reactions—with a focus on how model personality, falling costs, and surrounding tooling increasingly drive developer adoption.

Why this matters: Developers choosing between frontier models need to understand not just benchmark differences but cost trajectories and ecosystem tooling, which this episode directly addresses.

YouTube CEO Neal Mohan joins Big Technology to discuss how AI is reshaping content creation and discovery on YouTube, covering algorithmic creative control, AI-generated content moderation ("AI slop"), the Google DeepMind partnership, and how AI-powered answers might affect creator monetization.

Why this matters: YouTube's AI strategy and its DeepMind partnership signal how large platforms will reshape content discovery and creator economics in the AI era — directly affecting developers building on or for video platforms.

Eric Nguyen (CEO, Radical Numerics) dives into Genomic Language Models (GLMs) — how long-context architectures unlocked DNA-scale sequence modeling, how chain-of-thought reasoning applies to biological sequences, and why the same models that can synthesize functional viruses are also the best hope for biosecurity defense.

Why this matters: GLM architectures (long-context, chain-of-thought, multimodal) are converging with LLM techniques, meaning AI developers building foundation models will increasingly face biosecurity design decisions at the infrastructure level.

Flex-π is a 6B-parameter World-Action Model (WAM) for robotics that goes beyond RGB-only prediction by jointly denoising 3D pointmaps and DINO semantic features alongside color in a shared latent space, achieving 2–7× better performance over baselines on real-world bimanual manipulation tasks with greater demonstration efficiency.

Why this matters: The "free lunch" finding—that a frozen RGB video VAE encodes 3D pointmaps almost losslessly—has direct implications for researchers building robot policies without needing new sensors or pre-training pipelines.

This episode challenges the "tokenmaxxing" trend—where companies equate high token usage with AI progress—and makes the case for token efficiency as the smarter metric, offering practical guidance on how to measure it and avoid unnecessary cost pitfalls.

Why this matters: As AI inference costs scale with token usage, developers and engineering leaders need to optimize prompts and workflows for token efficiency rather than volume to keep production AI costs sustainable.

Replit CEO Amjad Masad and Andreessen Academy co-founder Gagan Biyani join a16z's Erik Torenberg to debate what education should look like in the AI era, covering project-based learning, the risks of over-professionalizing young founders, and how curiosity-driven exploration beats credential optimization.

Why this matters: Amjad's perspective on how AI reshapes the value of traditional credentials is directly relevant to developers and founders deciding how to invest in their own growth in the AI era.

A markets-focused episode covering oil price dynamics amid Trump's Iran threats, the sports betting regulatory landscape, and a brief editorial take dismissing the current AI debate — no deep technical AI content.

Reid Hoffman and Aria Finger unpack Dario Amodei's call to slow AI development, Anthropic researcher Evan Hubinger's >10% extinction risk estimate, and Sam Altman's support for slower development—while Reid argues for steering AI rather than pausing it, and outlines concrete risks like bioweapons, cyberattacks, and job disruption.

Why this matters: Frontier AI leaders publicly debating development pace and kill switches has direct implications for how AI products are regulated and shipped.

Economist Paul Kedrosky joins Ed Zitron to dissect how AI investment is distorting bond markets and US Treasuries, why AI data center economics don't add up, and concerning signals coming out of Anthropic.

Why this matters: If AI infrastructure economics are as broken as argued here, it has direct implications for the sustainability of the platforms and tools developers are building on top of.

Host Claire Vo runs a live blind benchmark pitting Claude Opus 5.5, GPT-6 Sol, GPT-6 Astra, and Luna against real work tasks—emails, PRDs, frontend prototypes, backend/agent work, SVGs, and video editing—then reveals which model won each category and where an LLM judge disagreed with her human scores.

Why this matters: Provides a practical, task-grounded comparison of frontier models (Opus 5.5 vs. GPT-6 Sol/Astra) across developer-relevant workloads like frontend prototyping, backend agents, and PRDs—useful for teams choosing a default model.

A short news roundup covering Anthropic's Opus 5.5 release, OpenAI's new GPT-6 models, and Meta's Muse and its growing traction in the AI space.

Why this matters: Developers building on Anthropic or OpenAI APIs will want to track the capabilities and positioning of Opus 5.5 and GPT-6 as they evaluate model upgrades.

Shane Thomas and Abhi Aiyer (Mastra CPO/CTO) dissect Jev, a viral "decision model" that returns calibrated probability outputs instead of text—claiming 20–200x speed and 40–400x cost improvements over LLMs—with Abhi live-demoing seven integrations into Mastra primitives including workflow branching, PII/abuse guardrails, tool-search pre-selection, model routing, and classifier-as-judge scoring. The honest verdict: a useful new classifier tool, not the revolution the hype promised.

Why this matters: Jev's RLCD-trained structured output approach offers developers a cheap, fast classifier primitive for guardrails, routing, and workflow branching—with concrete Mastra integration patterns shown live.

John Platt (inventor of SVM's SMO algorithm, sklearn's Platt scaling) joins Latent Space for a wide-ranging 2-hour conversation covering Google's ERA (Empirical Research Assistance) — an LLM-driven "auto-Kaggle" agent that uses Monte Carlo Tree Search-style optimization to solve any scoreable scientific problem — plus AI for climate change (contrail reduction), fusion, and quantum computing.

Why this matters: ERA's architecture (Gemini + UCB-guided notebook mutation tree) represents a concrete, open-sourced blueprint for building LLM-powered scientific optimization agents that developers can study and adapt.

Meta's Muse AI agent has topped the App Store charts, triggering a platform war as Amazon blocks it from shopping while Shopify welcomes it — raising urgent questions about who controls the customer relationship in an agentic commerce era. The episode also covers Grok 4.7, AI liability, and cross-lab safety testing.

Why this matters: The Amazon vs. Shopify split over AI agent access signals a coming platform fragmentation that will directly shape how developers build and integrate agentic commerce experiences.

A hands-on practitioner review of Claude Opus 5.5 after months of switching to Codex, covering real agentic task results, frontend prototyping, cost math, alignment behavior, and a frank model-stack comparison.

Why this matters: Opus 5.5's 40% cost reduction and revised alignment approach have direct implications for developers building agentic pipelines where per-token pricing compounds across long-running tasks.

Ben Horowitz and Gagan Biyani introduce the Horowitz and Andreessen Academy, a new project-based learning institution in San Francisco designed for aspiring builders in the AI era, arguing that hands-on doing—not traditional studying—is the right model when AI is rapidly reshaping the skills needed to create and work.

Why this matters: As AI compresses the learning curve for technical skills, alternative education models focused on real project-building could become a meaningful pipeline for the next generation of AI-era developers and founders.

A practical crash course distilling thousands of hours of LLM experience into 10 concrete best practices for getting the best outputs from any AI chatbot—covering ChatGPT, Gemini, Claude, and more as they increasingly converge in features.

Why this matters: As major LLMs converge in capabilities, a unified set of prompting best practices becomes actionable across tools, saving developers and power users significant trial-and-error time.

Brandon Purcell (Director of PM at Tiger Data) explains how TimescaleDB extends PostgreSQL with hypertables and Hypercore to handle time-series workloads at scale, covering zero-copy database forking for agent-based workflows and why standard relational databases struggle with continuous, append-heavy data streams.

Why this matters: Zero-copy database forking for agent-based workflows is a concrete capability developers building AI agents on Postgres-backed stacks should be aware of.