The Software Factory Explained

An article we liked from Thought Leader Alex Lieberman of Tenex:

What the hell is a Software Factory?

A non-technical leader's guide to The Software Factory, how it works, why it's in vogue, and how to get your company to operate one.

In 2023, everyone was hype about ChatGPT. In 2024, it was GenAI. 2025 was the year of Agents. And 2026 started with OpenClaw in vogue, but now attention has turned to The Software Factory.

Unless you're an engineer or take residence in the depths of X, you may not know what a Software Factory is or why you should care.

But when some companies are attributing 90% of their production software to AI (read: @AnthropicAI) and best-in-class ICs are matching the output of a 20-person pre-AI engineering org, you need to care.

Reality is, most companies are nowhere close to having a true Software Factory.

I run a company (@tenex_labs) where one of the things we now do is help enterprises build these. We set up the infrastructure and tooling, then run the assessment, training, and change management to get an engineering org ready for an entirely new way of building software. So I look at a lot of these orgs. And the gap between what CEOs think they have and what they actually have is enormous.

So let me break the whole thing down. What a software factory actually is, why it's suddenly everywhere, and a simple way to figure out exactly how close your org is. Even if you've never written a line of code in your life.

What is a Software Factory?

Strip the jargon and it's simple: a software factory is an engineering org where building software works less like craft and more like a production line.

In the craft model, a human sits down and writes the code by hand. Think of it like hand-building custom cars one at a time: flexible, but slow and inconsistent. Output scales with how many skilled people you can hire and how fast they can type. It's bespoke, slow, and most context lives in people's heads.

In the factory model, the work is industrialized. It's like a modern auto manufacturing plant: faster, more reliable, and scalable, but requires upfront investment in processes and tools. There's a repeatable line (write, review, test, deploy, monitor) and software moves through it with standardized steps, automated quality checks, and as little manual labor as possible. Humans design the line and handle exceptions. The machines do the reps.

The most extreme version I've seen comes from a system @simonw wrote about, built by a company called StrongDM. Two guiding principles:

That sounds insane. It's also the whole point. A factory isn't "humans coding faster with AI." It's a system where the humans move up a level, from doing the work to directing it.

Why is the term suddenly everywhere?

Three things hit at once.

1) The proof points got undeniable. @ryancarson now runs what he calls a "Code Factory." Agents write the code, review it, run the tests, triage the errors, and watch production, while he sets the guardrails. He's shipped over a thousand pull requests this way. Basically a one-person software company. Meanwhile Anthropic says 90%+ of its code is AI-written, Google says 75% of new code is AI-generated (up from 25% a year ago), and at OpenAI, 95% of engineers use their internal agent. The ones who lean in open 70% more pull requests than their peers (Lenny Rachitsky on X).

2) Big names are giving it oxygen. Microsoft is now pitching an "Agent Factory" as the new way software gets built. When the giants adopt the language, it goes mainstream fast. @chamath talks frequently about the Software Factory his team at 8090 is building. @garrytan hits on this factory concept with GStack, GBrain, and his newest article titled "Stop building Foxconn factories for your agents."

3) The metric became a flex. "X% of our code is written by AI" became the way CEOs signal they're ahead. Not the right metric to flex, but it earned agentic engineering attention nonetheless.

The framework: The Software Factory Ladder

In order to build a Software Factory, you need to be able to observe & orient yourself to where your engineering org is on the factory ladder.

Five levels, 0 through 4. You should be able to place your org in under a minute.

To make it concrete, I'll run the same scenario through every level: a customer hits a bug. The "Place Order" button stops working. Watch what shifts from human to AI as you climb.

Level 0. Artisan. A customer emails support. Eventually an engineer hears about it, reproduces it, writes the fix by hand, another engineer reads the code to review it, someone manually tests it, and a human deploys it. AI is nowhere. Everything is a human.

Level 1. Assisted. Same exact steps, same humans, except the engineer uses an AI copilot to…