Building a Software Factory that actually works (Full Course)

September 14, 2026

AI Summary

5 min read

Ross Mike, a developer building with AI agents, argues that the real bottleneck in shipping software isn't the model you use—it's your workflow. He calls this structured approach a "software factory," and it has nothing to do with buying a specific tool. It is, as he puts it, "completely harness and model agnostic." The factory is a set of markdown files that dictate how an agent should work, turning a chaotic back-and-forth into a repeatable assembly line.

The core problem: agents overstep and write sloppy code

Most people prompt an agent in a linear fashion: type a request, watch it build, tweak, and repeat. This works for a single feature, but it breaks down fast. Agents working on the same branch will overwrite each other's work—a common complaint on social media. "The agent will do what you tell it to do," Mike explains. If you ask it to update a landing page and also speed up API calls, it might delete design work while refactoring backend logic. The deeper issue is that even when code works, it is often poorly structured. Mike notes that models like GPT-5.6 can produce functional code that a reviewer like Fable will call "disgusting"—full of duplication and dead code. A factory solves both problems by enforcing isolation and quality standards before a human ever sees the result.

Step one: isolate to prevent conflicts

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

Never miss an episode of The Startup Ideas Podcast

Get every new episode summarized in your inbox — free, ~5 minutes to read.

No spam. Unsubscribe anytime.

What you'll learn

1 (00:03) **What a Software Factory Actually Is** - The core concept: using AI agents to ship high-quality software in a structured, assembly-line fashion, not just prompting iteratively.
2 (02:17) **The Core Workflow: Isolate** - The first step of the software factory: creating isolated work environments so agents don't overwrite each other's work.
3 (05:21) **How Isolation Works in Practice** - A walkthrough of the isolate step with a diagram, showing how it enables parallel development like a real engineering team.
4 (09:38) **Why Non-Technical Builders Need Isolation** - The host explains why this matters for the growing number of non-technical people building with AI.
5 (11:35) **Step Two: Build with Code Structure** - The second step gives agents guidelines on how to write quality, maintainable code, not just code that works.
6 (13:18) **Step Three: Prove** - The third step forces agents to provide visual or numerical proof that their work actually functions as intended.
7 (16:29) **Real Examples of Proof in Action** - The host pulls up actual PRs to demonstrate how the proof step works in practice.

+ Full timestamped outline available in the app

Guests on this episode

Ross Mike

Show Notes

Get Your Complete Financial OS at https://startup-ideas-pod.link/brex_SIP

I welcome Ras Mic back to the pod to explain the phrase "software factory." Mic shares his screen and walks through the exact system that he runs today. His factory has four steps: isolate, build, prove, and ship. He keeps the whole system in five or six markdown files, so it works with any model and any harness. By the end of this episode, you can boot up your own factory, run many agents in parallel, and trust the code that comes back.

Create your own Software Factory: https://startup-ideas-pod.link/ras-software-factory

Timestamps

00:00 – Intro

02:17 – Software Factory Definition

03:44 – Why the Software Factory Matters

05:23 – Step 1: Isolate With Git Work Trees

11:34 – Step 2: Build With the Code Structure Skill

14:48 – Step 3: Prove With Evidence-Driven Testing

22:25 – Step 4: Ship With Grep Loop and Greptile

26:52 – The Physical Factory Analogy

29:21 – A Software Factory Is Markdown Files

30:02 – Closing Thoughts

Key Points

A software factory is a workflow of skills and domain knowledge, so it runs with any model and any harness.
Isolate: every feature starts in a fresh git work tree branched from origin main, so each agent keeps its own station.
Build: a code structure skill makes the agent write service layer code that a human developer can read.
Prove: the agent records a before state and an after state as video, screenshots, or numbers.
Ship: Greptile scores the PR, and the agent loops back to build until it earns five out of five.
Mic runs up to 15 features in parallel and reviews the visual proof instead of the raw code.

The #1 tool to find startup ideas/trends - https://www.ideabrowser.com

LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products