#model-context-protocol
68 episodes
#5910: Revit MCP in a VM: What Actually Works
Autodesk's official Revit MCP server is locked down and reliable-ish. The community ones execute arbitrary code on an open port. Neither runs unatt...
#5799: Two Agents, One Approval Gate: Wiring a Dictation Workflow
A dictated prompt becomes an episode — if one agent resolves facts, a second tightens prose, and a human approves the dispatch.
#5771: MCP Connectors: Talk to Your Data vs. Open Data
Same protocol, two patterns: pointing agents at your own database, and pointing them at public data — with very different results.
#5770: Talking to Your Data: What MCP Doesn't Solve
MCP standardized the pipe. It has no opinion about what flows through it — and that's where the hard part lives.
#5606: Custom GPTs Retire: Skills, Plugins, and MCP Explained
Custom GPTs retire December 11. Here's how Skills, Plugins, and MCP servers actually replace them — and what doesn't migrate.
#5520: Keeping a Fine-Tune Alive Across Model Releases
Daniel wants to fine-tune DeepSeek Flash 4.1 on edited podcast scripts. The hard part isn't training — it's surviving the next release.
#5467: Small Models, Big Schemas: When JSON Constraints Backfire
Small models plus strict JSON schemas should be a safe bet. A 15,000-generation study found the opposite.
#5463: What Happens When You Click "Connect" on an AI Plugin
You click accept, the tool runs, and a credential you never saw is doing the work. Here's who actually owns it.
#5455: Custom GPTs Are Sunsetting: What Happens to Your Actions
OpenAI is retiring Custom GPTs — and custom Actions don't survive the migration. What that means for your workflow.
#5208: Why Tech Consortia Keep the Stack From Flying Apart
From GSM to HDMI to LTO, why competitors form standards bodies — and what happens when the commitments break.
#4713: MCP vs Browser-Use: The Agentic AI Showdown
Two patterns are fighting to define how AI agents interact with software. Which one wins shapes the next decade of building.
#4712: Codex vs Claude Code: The Real Agentic Coding Showdown
Benchmarks, model variants, and MCP setups — which coding agent actually wins where it matters?
#4703: MCP Binary Payloads: Why URLs Beat Base64
MCP can't carry raw bytes. Here's why the URL pattern is the only legal move—and what the File Uploads Working Group might change.
#4701: Why Claude Code Re-Introduces Itself to Your Repo
Claude Code remembers your preferences but not your codebase. We explore why that's by design and how to work around it.
#4700: Browser Agents: Three Paths to One Protocol
How screenshot watching, dev-tools extensions, and native MCP in Chrome are converging — and what wins.
#4663: Python vs TypeScript for MCP Servers: The Real Decision Rule
Both are first-class for MCP, but the real choice depends on what your server talks to. Here's the framework.
#4582: Designing a Home MCP Gateway for Conversational AI
How to securely connect ChatGPT to a home-hosted MCP gateway — transport, auth, and the missing pieces.
#4580: The Tiny AI Tools That Beat Big Agents
Why small, single-purpose AI utilities outperform autonomous agents for daily work — and what interface they need.
#4522: The Sixth Shim: When MCP Becomes a Middleware Factory
Six servers, six shims, no architecture. A taxonomy of MCP integration pain.
#4521: MCP Transport Architecture in 2026: Gateway Design
Three transports, one gateway. How to bridge stdio, legacy SSE, and streamable HTTP in production.
#4167: Why AI Can’t Read Your Gmail (Yet)
Your AI misses emails and fails to send replies. It’s not dumb—the pipes are leaky.
#3755: Hermes vs OpenClaw: Mobile-to-Server AI Frameworks
Why developers are leaving OpenClaw for Hermes—and why mobile-to-server AI interaction remains unsolved.
#3271: LLMs as Parsers, Not Calculators
Stop letting LLMs do math. Use them to parse messy text, then let deterministic code handle the numbers.
#2690: Where Agent Builders Actually Gather
The MCP community, A2A protocol, and Linux Foundation are building the professional identity of agentic AI right now.