Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

AI coding is blurring the line between software user and builder

When the cost of AI code is almost zero, that changes who participates.

[Photo: Getty Images]

Profile photo of Louis Camassa

Director of Product · Rithum

For decades, building software required significant cost. You needed people and process. You needed tools, infrastructure, sprint cycles, coordination plus capital, structure, and distribution plans.

The fewest people I ever launched a minimum viable product with was three: me as the idea person, an engineer, and a designer. To move faster, you added more people.

That model is collapsing.

AI coding tools like Claude Code are doing for software what Spotify did for musicians and YouTube did for creators. They are moving the means of production from institutions to individuals.

And when that shift happens, the world reorganizes.

ELIMINATE THE GATEKEEPERS

Before Spotify, music required studios, labels, and physical distribution. Shelf space was scarce. A small group of executives decided who was worthy of being heard.

My friend Bill Manspeaker, aka Hollywood Bill, founder of the Grammy-nominated punk-metal band Green Jelly, once explained how to hit the music billboards. You broke a record in smaller markets first, created momentum, then climbed toward national exposure.

When Spotify removed the gatekeeper, the bottleneck disappeared. Anyone could upload. Today more than 100,000 tracks are uploaded every day.

YouTube and TikTok repeated the pattern. Cameras moved into phones. Editing moved into apps. Distribution became seamless.

The shift was not just more content. It was a change in power.

Value moved away from production rigor and toward discovery. In a world of infinite supply, the curator, the algorithm, and the platform gained leverage.

AI CODING IS ELIMINATING BARRIERS

For most of modern history, software belonged to specialists. The barrier was capital. Now AI coding is at the same inflection point as music was 20 years ago.

Fifteen years ago, I was building custom e-commerce sites on top of a base platform. We were not coding from scratch, but this was pre-Shopify. It still required people and effort.

Across 25 projects, we generated about $714,000 in revenue with roughly a 47% profit margin. The typical build involved about eight contributors across product, engineering, design, QA, marketing, and support. It ran around 380 hours and took about 18 weeks. And those 18 weeks were a grind.

AI coding tools are compressing that barrier dramatically.

Anthropic’s Super Bowl commercial pushed me to try Claude Cowork. I was an early paid ChatGPT user, but the Claude Cowork buzz and FOMO got to me. I spent about four hours across three mornings and built a health dashboard that logs into Garmin Connect, pulls my daily watch data, and emails me a report every morning.

What amazed me is Garmin does not offer a clean public API for this use case. Claude found an open-source workaround I did not know existed. I installed what it recommended, ran it locally, and when things broke, pasted logs into Claude. It rewrote the script until it worked. I essentially copied and pasted, like following a Blue Apron recipe.

Within about 70 minutes, I had a live system pulling steps, heart rate variability, sleep, stress, and hydration, with the ability to format it cleanly and send it automatically. I pushed it to GitHub, deployed it to Railway for $20 a month, and when Gmail would not cooperate, Claude suggested Resend, fixing that problem too.

No engineers. No designer. Just prompts and persistence.

The cost of software creation is falling toward zero. That changes who participates.

THE RISE OF THE MICRO BUILDER

Spotify and YouTube did not just create superstars. They created a middle class of creators,  niche podcasters, independent musicians, and one person media companies with millions of followers.

AI coding tools will do the same for software.

We are about to see the rise of the micro builder—individuals launching highly specific tools that serve narrow audiences. Like a scheduling app for pickleball players trying to find pickup games in small towns.

In the old world, that market was too small to justify the cost. In the new world, it is viable.

This is the long tail of software. But the real winners are the aggregators.

Spotify empowered musicians, but it captured listening. YouTube empowered creators, but it became the discovery engine.

In AI coding, the aggregators will likely be cloud providers, model platforms, app stores, and distribution layers that sit between builders and users.

LOWER BARRIER WITH HIGHER COMPETITION

There is a cost to democratization. When anyone can create, everyone does. As content exploded, attention fragmented. We tell ourselves the best will rise to the top. In reality, the most optimized often wins.

Software is heading in that direction. When anyone can ship an app, the constraint becomes distribution, brand, and trust. The ability to build becomes commoditized. The ability to be discovered becomes strategic.

Craft does not disappear, but its premium shifts.

SOCIETY WILL SHIFT TOO

When music and media democratized, career paths changed. Teenagers became entrepreneurs. The line between consumer and creator blurred.

AI coding will blur the line between user and builder.

Inside my company, employees are building internal tools without waiting for IT. Consultants are deploying apps instead of slide decks. Student entrepreneurs I mentor are building prototypes instead of writing essays.

We are trading centralized quality control for mass participation. The landscape will be noisier and more chaotic. It will also be more innovative.

Spotify did not kill music. It changed who could make it. YouTube did not kill video. It changed who could publish it.

AI coding will not kill software engineering. It will change who can ship.

The gatekeeper era of software is ending. The attention economy of code is just beginning.

Louis Camassa is the director of product for Rithum.