A.I. Joe

Blog

Everything we publish, in one place.

Practical writing on AI-assisted development, structured agentic coding, and keeping humans in control of the systems they build.

The archive

Sep 14, 2026

When Green Dashboards Hide Broken Systems

Sep 13, 2026

The Benchmark Lied by Omission: Why AI Speedups Break on Real Traffic

Sep 12, 2026

AI Code Doesn't Break Where You're Watching

Sep 11, 2026

AI Writes Steps. Production Needs States.

Sep 10, 2026

The Silent Bug Is Always More Expensive Than the Loud One

Sep 9, 2026

The Bugs AI Ships Are the Ones Nobody Ever Wrote Down

Sep 8, 2026

The Silent Feature: Why Working Code Still Leaves You Blind at 3am

Sep 7, 2026

The Hidden Web Behind Every AI Commit

Sep 6, 2026

When AI Solves Twelve Problems You Only Have One Of

Sep 5, 2026

The Bugs That Live Between Correct Files

Sep 4, 2026

AI-Written Workers Quietly Assume Only One of Them Exists

Sep 3, 2026

The Real Reason AI Attribution Feels Risky in Code Review

Sep 2, 2026

When Clean Names Lie About What Runs Underneath

Sep 1, 2026

The Best Code Review Question Is 'What Can We Cut?'

Aug 31, 2026

Vague Prompts Are Why Your Code Keeps Growing

Aug 30, 2026

The Gap Where Two Correct Functions Disagree

Aug 29, 2026

AI Code Looks Finished. It Just Hides Its Reasoning

Aug 28, 2026

AI Assistants Don't See Your Service Boundaries, They See File Paths

Aug 27, 2026

The Real Reason Clean AI Code Still Scares Engineers

Aug 26, 2026

Why Typing Less Actually Slows Down AI Coding

Aug 25, 2026

The Math Behind Detailed AI Prompts Beating Quick Ones

Aug 24, 2026

The Shortcut in Your First AI Prompt Never Goes Away

Aug 24, 2026

The Design Reasoning Behind Your Code Disappears Between Sessions

Aug 23, 2026

The Bug That Hides in the Handshake Between Systems

Aug 22, 2026

Your Skill Library Is Quietly Sabotaging Claude Code

Aug 21, 2026

When Tests Just Echo the Code That Wrote Them

Aug 20, 2026

The New Bottleneck: Checking AI's Work, Not Writing It

Aug 19, 2026

The Bottleneck Was Never the Typing

Aug 18, 2026

The Silent Failure Mode in AI-Generated Architecture

Aug 17, 2026

When Every Commit Is Reasonable But the System Isn't

Aug 16, 2026

When AI Code Passes Every Test but Answers the Wrong Question

Aug 15, 2026

The Silent Version Gap That Sinks Working Code

Aug 14, 2026

The Prompt Length Trap Senior Engineers Fall Into

Aug 13, 2026

AI Writes Code Fast; It Erases the Reasoning Just as Fast

Aug 12, 2026

The Reasoning Behind Your AI Code Vanishes the Moment You Approve It

Aug 11, 2026

The Real Danger Is the AI Code That Works

Aug 10, 2026

Vague Names Become Design Decisions When AI Writes the Code Around Them

Aug 9, 2026

The Bug Isn't in the Code, It's in What the Model Never Saw

Aug 8, 2026

Every Commit Is Correct and the Repo Still Rots

Aug 7, 2026

The Bugs AI Writes Are the Ones Nobody Wrote Down

Aug 6, 2026

When More Prompt Detail Backfires on AI Code

Aug 5, 2026

The Bug Isn't in the Code You Changed

Aug 4, 2026

The Real Sandbox Escape Route Is Content, Not Commands

Aug 3, 2026

The Race Conditions AI Never Sees Coming

Aug 2, 2026

AI Coding Assistants Default to the Tutorial Answer, Not the Wise One

Aug 1, 2026

The Import Line Nobody Reviews

Jul 31, 2026

The Hidden Reason AI Writes Buggy Retry Logic

Jul 30, 2026

The Review Instinct That Betrays You With Clean Code

Jul 29, 2026

Three Names, One Concept: The Search That Fails at Midnight

Jul 28, 2026

The 80% Statistic Is Measuring the Wrong Thing

Jul 27, 2026

AI Can't Tell You When It's Guessing

Jul 26, 2026

When the Bug Isn't in the Code, It's in the Ask

Jul 25, 2026

The Hidden Bill for Fast AI Code

Jul 24, 2026

Hallucinations Aren't a Prompting Bug, They're a Missing Layer

Jul 23, 2026

AI Code Is Confident, Not Correct

Jul 22, 2026

When AI Hands You a Framework You Didn't Ask For

Jul 21, 2026

Your AI Assistant Is Padding Your Codebase With Popularity Contests

Jul 20, 2026

The Over-Detailed Spec Is Silently Sabotaging Your AI Code

Jul 19, 2026

The Snap You Didn't See Coming: How Fast Code Generation Hides Structural Debt

Jul 18, 2026

The Hidden Cost of AI's Favorite Interface

Jul 17, 2026

Same Prompt, Different Code: The Model Isn't Guessing, You Are

Jul 16, 2026

AI Code Works Perfectly Until the Requirement It Wasn't Told About Arrives

Jul 15, 2026

The Real Cost of AI Code Isn't Speed, It's Amnesia

Jul 14, 2026

AI Code Ages Faster Than You Think, and the Bill Comes Due at Refactor Time

Jul 13, 2026

The Real Reason AI-Written Bugs Slip Past Review

Jul 12, 2026

The Bug Isn't in the Code, It's in the Unstated Assumption

Jul 11, 2026

The New Bottleneck Isn't Typing, It's Deciding What to Delegate

Jul 10, 2026

The Backdoor Isn't in the Code, It's in the Text

Jul 9, 2026

Context Windows Fail as Storage, Succeed as Budgets

Jul 8, 2026

When Your AI Tools Disagree, Don't Rush to Fix It

Jul 7, 2026

The Assumptions Buried Inside Tidy AI Code

Jul 6, 2026

The Test That Passes and Still Lies to You

Jul 5, 2026

AI Removed the Cost That Kept Over-Engineering in Check

Jul 4, 2026

Why AI Never Sounds Unsure, Even When It's Wrong

Jul 3, 2026

Your Commit Log Just Lost Its Ghostwriter

Jul 2, 2026

Fast Code, Slow Debt: The Hidden Tax of AI-Generated Software

Jul 1, 2026

Passing Tests Aren't the Same Thing as Good Code

Jun 30, 2026

Prompts Are Code, Not Config. Version Them Like It.

Jun 29, 2026

A Kinder Wording Took Down Production, and No One Touched the Code

Jun 28, 2026

Your Prompts Fail Like Bad Syntax, Not Bad Delegation

Jun 27, 2026

When Prompts Rot Without Anyone Noticing

Jun 26, 2026

Your Codebase Is Already Talking to the AI, Prompts Just Interrupt

Jun 25, 2026

Solo Developers Now Ship Global Products Because the Support Staff Became Software

Jun 24, 2026

The Liability Never Transfers to the Model

Jun 23, 2026

The Bugs That Pass Every Test Anyway

Jun 22, 2026

The Question That Separates Engineers From Prompt Operators

Jun 21, 2026

The One-Person Software Company Just Became Viable

Jun 20, 2026

The Real AI Moat Isn't the Model, It's the Plumbing

Jun 19, 2026

Your Inference Bill Was Decided in a Fab Eighteen Months Ago

Jun 18, 2026

Judgment, Not Prompts, Decides Whether AI Code Ships

Jun 17, 2026

The Wrapper Test Every AI Founder Should Run Before the Next Model Release

Jun 16, 2026

AI Didn't Kill Specialization. It Made Crossing the Stack Cheap Enough to Matter

Jun 15, 2026

The AI Funding Flood Is a Free Lab, Not a Mandate to Buy

Jun 14, 2026

AI Companions Move Judgment, Not Replace It

Jun 13, 2026

When an AI Agent Refactors Your Codebase, Who's Actually Responsible

Jun 12, 2026

Your Staging Tests Are Lying to You About Production Quality

Jun 11, 2026

Senior Engineers Now Spend Their Time Judging Code, Not Writing It

Jun 10, 2026

When the Agent Grades Its Own Homework

Jun 9, 2026

Your Real Bottleneck Isn't the Model, It's Your Own Head

Jun 8, 2026

When the Tests Go Green but the Bug Is Still There

Jun 7, 2026

Leaderboard Rank Doesn't Predict Real-World Coding Performance

Jun 6, 2026

The Hidden Ceiling Under Every AI Model

Jun 5, 2026

AI Ships Into a Trust Deficit It Did Not Create

Jun 4, 2026

Why Ethics Statements Always Arrive After Launch Day

Jun 3, 2026

The Real Competition Moved Up the Stack

Jun 2, 2026

The Middle Market Software Never Bothered To Serve

Jun 1, 2026

The Junior Hiring Freeze Is a Ten-Year Time Bomb

May 31, 2026

Localhost Confidence Meets Production Reality

May 30, 2026

Gold Medal Math, Failing Grade on Clocks: What AI's Uneven Skill Map Means for Builders

May 29, 2026

When AI Talks Back, Bring Your Messy Thoughts, Not Your Polished Questions

May 28, 2026

Code Got Cheap. Judgment Didn't.

May 27, 2026

Agentic AI Doesn't Add Uncertainty, It Compounds It

May 27, 2026

Why Enterprise Agents Fail After the Demo Succeeds

May 26, 2026

When Typing Stops Being the Bottleneck, Judgment Becomes the Product

May 24, 2026

Coding Capacity Just Stopped Being Scarce

May 22, 2026

The Plumbing Behind the Model Is the Real Product

May 21, 2026

Why Knowing the Problem Beats Knowing the Code

May 20, 2026

AI Code Never Sounds Unsure, Even When It's Wrong

May 19, 2026

When Code Gets Cheap, Thinking Becomes the Scarce Resource

May 18, 2026

Your First AI-Built System Rewires How You Debug Everything After

May 15, 2026

Good Structure Lets AI Help; Bad Structure Lets AI Hurt

May 14, 2026

The Real Bottleneck Isn't Code Anymore, It's Judgment

May 13, 2026

Why Scar Tissue Beats Speed in AI Coding

May 12, 2026

The Bottleneck Moved From Frameworks To Thinking

May 12, 2026

Why AI Coding Sessions Quietly Drift Off Course

May 11, 2026

Fast Guessers Need Guardrails, Not Faith

May 8, 2026

The Hidden Layer Between a Working Demo and a Working Product

May 7, 2026

The Knowledge Gap AI Coding Leaves Behind

May 6, 2026

When Code Gets Cheap, Judgment Gets Expensive

May 5, 2026

Confident Code Hides Its Own Bugs

May 4, 2026

Where AI Optimizes the Wrong Problem

May 3, 2026

Why Your AI Prototype Falls Apart the Moment It Meets Users

May 1, 2026

Coherent Pieces, Incoherent Whole

Apr 30, 2026

AI Removed the Friction That Used to Teach You to Code

Apr 29, 2026

AI Made Coding Cheap. It Didn't Make a Business Cheap.

Apr 29, 2026

Reroll Culture Is Killing Your Codebase

Apr 29, 2026

When Your Codebase Forgets Why It Was Built That Way

Apr 26, 2026

Your Agent Framework Is a Chassis, Not a Convenience

Apr 25, 2026

Who Holds the System Together When AI Writes the Parts

Apr 24, 2026

When Passing Tests Hide a Broken Design

Apr 23, 2026

The Model Is Furniture, Not the Foundation

Apr 22, 2026

When Building Gets Cheap, Judgment Gets Expensive

Apr 21, 2026

When Code Outpaces Understanding

Apr 21, 2026

Why Niche Trades Are Suddenly Worth Building Software For

Apr 20, 2026

Opus 4.7's Real Upgrade Is Knowing When to Stop Guessing

Apr 16, 2026

Fast Code Is Not the Same as Good Code

Apr 15, 2026

The Demo That Kills Your Own Project

Apr 14, 2026

AI Coding Speed Depends on Decisions Made Before You Prompt

Apr 13, 2026

Domain Knowledge Just Became the Scarcest Resource in Software

Apr 12, 2026

AI Code Ships Fine Until Nobody Can Say Why It Changed

Apr 12, 2026

When AI Writes the Code, Who Owns the Design

Apr 10, 2026

When AI Writes Faster Than You Can Understand

Apr 8, 2026

AI Doesn't Need a Better Prompt. It Needs a Governor

Apr 7, 2026

Three AI Tools Agreeing Doesn't Mean They're Right

Apr 6, 2026

Why Fast AI Code Generation Creates Slow Debugging Later

Apr 5, 2026

Why Domain Experts Are Beating AI Labs to the Punch

Apr 4, 2026

AI Orchestration Layers Explained

Apr 3, 2026

Architecture Is the Bottleneck AI Cannot Skip

Apr 3, 2026

The Silent Drift Inside Automated Systems

Apr 3, 2026

Fast Code Can Hide a Slow-Motion Wrong Turn

Apr 2, 2026

The Return of the Software Architect

Apr 2, 2026

The Over-Automation Trap

Apr 1, 2026

The Pause That Protects Your Architecture

Mar 31, 2026

The Automation Stack of the Future

Mar 30, 2026

Experience as a Superpower in the AI Era

Mar 29, 2026

Why Small Teams Will Beat Large Software Companies

Mar 29, 2026

When AI Refactors Break Your Architecture

Mar 28, 2026

AI Agents vs Traditional Software

Mar 27, 2026

Stop Feeding Claude Code Small Questions

Mar 27, 2026

Why Coding With AI Makes Expertise Worth More, Not Less

Mar 26, 2026

The Architect–Engineer Hybrid

Mar 26, 2026

Claude Code in Practice: Extended Context

Mar 25, 2026

The Danger of “Prompt and Pray” Development

Mar 24, 2026

The Verification Principle: How Boris Cherny and the Chief Engineer Arrived at the Same Idea

Mar 23, 2026

Builder Story: Building a Podcast Factory With AI

Mar 22, 2026

The New Economics of Building Software

Mar 21, 2026

Claude Code vs Traditional Development

Mar 20, 2026

Why Domain Experts Now Have an Advantage

Mar 19, 2026

The Illusion of Working Code

Mar 18, 2026

Why AI Needs Architects, Not Just Programmers

Mar 18, 2026

The Rise of the One-Person Software Company

Mar 18, 2026

The Human Chief Engineer

Mar 18, 2026

The Biggest Mistake People Make With AI Coding

Mar 18, 2026

The Architecture Layer of AI Development

Mar 18, 2026

From Prompting to Engineering

Mar 18, 2026

Designing Systems AI Can Safely Modify