Insights for great software & great teams

How-to guides, tooling tips and real-world advice for developers and product managers tired of broken software. Let’s fix this mess together.

Sign up and get insights delivered to your inbox.

How-to guides, tooling tips and real-world advice for developers and product managers tired of broken software. Let’s fix this mess together.

Want to talk shop or tackle a problem in real time? Unpack it with a developer or product manager.

Get our newsletter

Tradeoffs chart for changeability, cost, and reliability

What is legacy software in the age of AI?

Software becomes legacy by succeeding long enough to accumulate the weight of every decision and shortcut made along the way. Users feel it as friction. Engineers feel their momentum slip. The business watches costs climb. AI changes two of those. The third may be getting worse.

Todd Kaufman

Legacy modernization

Three people emerge from darkness toward light and a neon green path

In the age of AI we’re in this together

River Lynn Bailey

River Lynn Bailey

Cognitive hazard study diagram showing a laptop, coffee mug, and documentation sinking into a tarpit

Organizational observability: The AI alignment problem starts upstream

Dave Mosher

$Puzzled math meme with a rubber duck wearing a fedora and trench coat$

The shockingly simple math for why we need test automation

Gabriel Côté-Carrier

Gabriel Côté-Carrier

Double Takes

+ Expand posts

- Collapse POSTS

Double Takes

TDDing with Claude is like having a pair that removes all the failing assertions every time you go to the bathroom.

Steve Jackson

State machines are the answer for those seeking determinism in the era of AI agents.

Dave Mosher

A Slack conversation left Doc reflecting on what bug bashes say about the systems we work in.

Doc Norton

Test-driven development is to programming what double-entry bookkeeping is to accounting.

Neal Lindsay

The fundamental character flaw of the software creator is not the misguided belief that they can solve any problem with software. Instead, it is the belief they can understand how their own software works before they create it.

Steve Jackson

We may need a SaaS appendix for Zawinski's law:

Any SaaS platform attempts to expand until it is a social network.

Matt Jones

Nesting boxes unfold from decorative to scaffolding to hardware

We're about to unwind fifty years of "Progress"

Every programming language, framework, and abstraction we've built exists to solve the same problem: humans have limited cognitive bandwidth. As agents become the primary authors of code, the rationale for all that scaffolding starts to weaken.

Doc Norton

ai-ml

quality

efficient software

process

systems

Rubber duck in a trench coat and fedora next to a toy robot

Field Report: All Things AI 2026

Test Double double agents provide a field report recap of the amazing 2026 All Things AI conference.

Cathy Colliver

ai-ml

teams

communication

efficient software

product-innovation

Desk in a home office shows one empty chair with a jacket and nearby coffee cup and notebook, and a second empty chair in front of a keyboard and computer monitor

What does "Good Code" even mean now?

For decades, we've optimized code for human readers. Agentic coding is forcing a renegotiation of what "craft" means, and that's shifting emphasis from code-level readability to system-level observability and problem-space incrementalism.

Doc Norton

ai-ml

quality

Laptop in a home office with screen displaying: Red, Green, Refactor

Red-Green-Refactor your context

You committed the fix. But everything you learned along the way went nowhere. A practice borrowed from TDD's red-green-refactor loop gives that hard-won knowledge somewhere to land.

Rick Reilly

ai-ml

efficient software

legacy-modernization

process

productivity

Roiling ocean

This has happened before. It's happening again.

Doc Norton has navigated two genuine sea changes in software development—mainframes to PCs and desktop to web. Agentic coding is the third, but unlike the prior shifts that expanded the profession, this one compresses it. The timeline to move is shorter than you think.

Doc Norton

ai-ml

professional-growth

community

efficient software

Toy race car headed for a smashed coffee cup and clock

Three Amigos with AI: Stop building the wrong thing faster

AI coding assistants are remarkably fast at building the wrong thing. This post adapts the Three Amigos practice to give AI an unambiguous contract—before the first commit, not after the fifty-seventh.

Andy Vida

ai-ml

process

productivity

efficient software

product-innovation

Line drawing of a rocky mountain peak with a base camp and people building structures

Quality you can’t generate: AI is only as good as your constraints

AI changed the cost structure of software. It didn't change the value structure. The value is no longer the code you write. It's the taste, judgment, and constraints you encode into the system that shapes what AI produces.

Dave Mosher

ai-ml

quality

efficient software

productivity

architecture

A figure looks off towards the horizon. On the left is a giant robot, on the right a splash of vibrant color, connected nodes and a city unfurl.

Why we're not chasing the AI hype (And what we're doing instead)

We want clients and prospective clients to know they can entrust us to solve problems with AI while remaining true to who we are, how we work, and the value we actually provide.

Todd Kaufman

teams

communication

founders

culture

strategy

You're N.E.A.T.!

Technology is a means to an end: answers to very human questions. We want to connect with other folks who share that belief.

Ways to connect

Test Double Executive Leadership Team

Learn about our team

Like what we have to say about building great software and great teams?