
Featured video
Voice AI in production - does it work?
A demo of several popular voice AI providers, how it works, and my experience with real customers talking to my agents.

Featured video
A demo of several popular voice AI providers, how it works, and my experience with real customers talking to my agents.
The journal

Learn how to get your products to show in ChatGPT or Claude using MCP, and how connectors like UCP and ACP work.

Understand how tools and MCP work in your customer facing AI agents so you can reduce hallucinations and get them to follow instructions better.

Jev is a new classifier model released by the co-creator of ChatGPT. It's not an LLM, but it's really fast and cheap way to make decisions using structured output. You should definitely use it in your apps, but it may not work the way you think.

Scott and Wes discuss Amp Orbs, the new Jev model, Grok Bot, slop grenades, and the whiteboard test for AI engineers.

Your customer facing AI agent prompts need 5 things: job, truth, workflow, guardrails, and FAQ. Avoid these 2 big mistakes: too much and too little.

An AI Agent is a loop with LLM calls, tools, and more, here's what that actually looks like.

Wes and Scott discuss LLM-assisted coding in cloud containers like Amp Orbs, and Wes shares an AI-built side hustle.

Wes and Scott discuss whether agentic coding atrophies your skills, demo a Codex workflow with computer use, and share news about the SWE job market.

Scott and Wes discuss how they use Codex and Claude in production at work, from workflow improvements to AI harnesses and automation.

Wes and Scott explore practical AI integration, misconceptions about AI progress, and the enduring value of creativity and strategic thinking.

Scott Bolinger and Wes Cole discuss AI, Skills, Open Claw, Codex, Pi, and what makes software engineers valuable beyond writing code.

Scott and Wes share their AI predictions for 2026, covering agentic coding, developer tooling, jobs, and interviewing.

Wes and Scott discuss how AI feels at work, why it isn't taking their jobs any time soon, and Google's Universal Commerce Protocol.
Even if AI writes all the code.

A discussion of coding agents, software engineers' value and jobs, missing hands-on coding, Anthropic agent skills, and Tailwind's business decline.

Scott and Wes discuss AI-built apps, lower barriers to MVPs, and why engineers need to focus on business value as software development changes.

Guest Richard Clayton, director of engineering in Care at GoDaddy, joins Scott and Wes to discuss AI in customer care and internal engineering.

Scott and Wes talk about Gemini 3 and Opus 4.5, then dive into Cloudflare workers and durable objects.

A short clip about LLM context windows from episode 6 of You're Absolutely Right.
The basics of building with LLMs

Wes and Scott discuss context windows, prompt scope creep, context overflow, MCP, and getting better model outputs.

Scott and Wes discuss testing multi-agent systems, Convex.dev, voice AI tools, and measuring AI's impact on engineering.

Wes and Scott discuss layoffs, a vibe-coding app, data models, Cursor's Composer model and new UI, and 0github.com.
A simple way to think about what AI agents are, what they do, and why getting them to work reliably is hard.
Building a demo app with an LLM is easy, because you don’t need it to work perfectly.
From my experience using Open AI to build an AI Assistant, hopefully it’s helpful!