Artificial Intelligence
Episodes whose main subject is AI itself: building with, training, evaluating, or debating AI systems. AI used only as a tool for another topic goes to that topic.
#4504: When Cars Took Over: Traffic Engineering's Origin Story
How 1920s traffic engineers borrowed from fluid dynamics, and whether AI routing could be a public utility.
#4497: Voice-First Stack Consolidation
Two transcription apps, two text expanders, no sync — and snippets about to fight each other. The path forward.
#4495: The Juice Vendor's Secret: Why Listing Works
Why a rapid-fire list of specific juices sells more than "fruit juices" — and how to apply it.
#4477: Structuring Raw Voice Prompts at Scale
How to decompose messy voice dictation into clean, routable components for AI workflows.
#4471: What's Actually Behind the API Endpoint?
How OpenAI, Stripe, and Netflix actually manage APIs at millions of requests per second.
#4469: The Missing Protocol for AI Software Development
No one has built a software development protocol designed for AI agents. Here's what that would require.
#4466: Free Your Shopping Cart: Dot Cart Explained
A new open file format lets you export, share, and compare shopping carts across any vendor.
#4465: True Crime Docuseries That Don't Earn Their Runtime
A countdown of docuseries that stretch thin stories across too many episodes — and the padding tricks they use.
#4462: How AI Imagines Your Skeleton
From image to fully rigged 3D model — the six computational stages that turn photos into animated characters.
#4458: How Headroom Compresses Agent Context Without Breaking Outputs
A deep dive into Headroom's content-aware compression pipeline that promises 20-95% token savings without changing agent answers.
#4457: OpenViking: A Filesystem for AI Agent Memory
OpenViking unifies memory, RAG, and skills into a browsable filesystem for AI agents — no vector database abstraction needed.
#4456: Inside the Podcast Pipeline: How 15 Weekly Episodes Get Made
From prompt to published episode — a full walkthrough of the automated production system running 15 shows weekly.
#4453: Workers vs Workflows: Recovering From SSD Failure
When your VPS SSD dies at 3AM, how do you know what broke? A practical guide to self-hosted worker management.
#4451: Python 3.14: Deferred Annotations, t-Strings & Free-Threaded Mode
Deferred annotations, t-strings, subinterpreters, and free-threaded mode — what actually changed in Python 3.14.
#4450: Intercepting Claude Code: The Proxy Ecosystem
How middleware proxies intercept, reroute, and transform AI tool traffic — and what it means for users.
#4449: Blur Before You Publish: Image Redaction Tools
How to automatically find and blur QR codes, credit cards, and sensitive text in photos before publishing.
#4448: The MK Claude Project: AI That Writes Political Manifestos
One citizen built a multi-agent AI to research, stress-test, and synthesize policy positions. The result? A forkable manifesto.
#4446: What Claude Fable 5's Suspension Reveals About AI Regulation
Three days after launch, the US government suspended Anthropic's Claude Fable 5. Here's what happened and why it matters.
#4445: AI Agents Keep Undoing Each Other's Work
Why AI coding agents silently revert each other's fixes — and three memory strategies that prevent it.
#4444: Testing the Unpredictable: QA for Agentic AI
How QA adapts when your AI system gives different answers to the same question every time.