Part 1: Foundations -- available now

The bottleneck isn't code anymore.

It's knowing what to build.

The definitive playbook for building software products with AI agents — written by a CTO who's done it. For vibecoders and CTOs alike.

Free to read. Creative Commons (CC BY).

What you'll learn

Five ideas that will change how you build software

These aren't AI hype takes. They're frameworks forged from shipping real products with agentic workflows.

01

The Renaissance Developer

Breadth over depth. The new competitive advantage is being good enough at everything.

02

The Acceleration Paradox

When coding gets 5-10x faster, five new bottlenecks emerge that nobody talks about.

03

The Digestibility Principle

Human working memory and AI context windows have the same constraints. Design for both.

04

Specification-Driven Development

The quality of your descriptions determines the quality of your product.

05

The "Good Enough at Everything" Principle

Why 70% competency across 5 domains beats 95% in one.

What readers are saying

Finally, a book that treats AI coding as a real engineering discipline, not a party trick.
Coming soon
The Renaissance Developer chapter changed how I think about my role as a senior engineer.
Coming soon
Practical, opinionated, and exactly the playbook my team needed to adopt agentic workflows.
Coming soon

Table of Contents

Four parts. One playbook. Idea to production in six weeks.

01

Foundations

Available

The Renaissance Developer mindset, what agentic coding really means, and first principles of software architecture.

02

The Playbook

Soon

The complete 6-week journey: idea, brainstorm, brief, requirements, design, implementation, testing, and deployment.

03

Patterns & Tools

Soon

Architecture patterns, interface design, testing strategies, and the agent-friendly toolchain.

04

Complete Example

Soon

A real, non-trivial product built using the entire playbook -- proving the 6-week promise.

Get notified when Part 2 drops

No spam. Just a single email when new chapters go live.

AI can write code faster than you can type. The question isn't whether to use it — it's whether you know how to direct it.

Start Reading Part 1