4 Week MVP · For people who can’t code
Turn your idea into a real, deployed app in four weeks.
Without learning to code. Without hiring anyone. Without spending another year thinking about it.
I went from barely being able to connect a printer to my computer to building a real software company with AI — over 100 users in the first four weeks, and $100,000 in written investment offers within two meetings of launching.
Those are my results, and they are not typical. I have no idea what you'll do — that depends on your idea, your market, and how hard you work.
You are four weeks and one system away from owning real software, with real users.
Not a prototype. Not just a file or a folder. A live application that strangers can sign up for and use — built by you, without writing code, without hiring anyone.
You already know AI can build software. That’s not the problem. You’ve probably even tried — opened the tool, typed the idea, watched something appear, then hit a wall of red text and quietly closed the laptop.
Or worse, you never got that far. You’re still:
- Refining the idea instead of building a version of it
- Waiting until you understand databases well enough to start
- Keeping it quiet in case someone takes it
- Reading one more thread about which tool is best
- Certain that the plan needs one more pass before it's ready
None of that is stupidity. All of it feels like progress. That is precisely why it’s expensive — thinking produces the same sensation as working and none of the artifacts.
Most entrepreneurs sit around thinking about their business and never actually do anything to it, or about it.
The reason this exists
A year ago, my friend Bill and I made a deal.
We’d both been fired up about what AI had just made possible. He had good ideas — genuinely good ones. We agreed we’d each go execute and race each other.
A few weeks ago he called to congratulate me on my company. I asked how his was going. Long pause. Deep sigh.
“Well — I’m still thinking about it. I was able to get a landing page up.”
Bill isn’t lazy and he isn’t stupid. He’s one of the sharper people I know. He spent a year on it — researching, refining, reconsidering, opening his laptop many, many times.
He just never built anything. And that landing page is the most expensive thing either of us made that year. Not because it cost money. Because it cost him the year.
Why most of this category fails you
Four things decide whether you actually finish.
Every course in this space raises the first one and ignores the other three. That’s why you’ve bought things before and not finished them. This is built the other way round.
Something real, on the internet, with your name on it
Not a certificate. Not a folder of prompts. A live URL you can send to your mother, and a repository with your commits in it. The thing you've been describing at dinner parties for three years, existing.
Every known way to fail, already handled
The Error Atlas decodes the errors you'll actually hit. Fifty pre-scoped ideas so you can't pick something unfinishable. A complexity rubric that scores your idea before you waste a week on it. The 20 mistakes, each with the fix.
A live URL on day one. A working app in four weeks.
Not "eventually." Module 0 ends with something deployed to the internet in ninety minutes — before you've built anything. You get the win first, then the product.
You are not learning to code. At all.
Thirteen fill-in templates. Tagged starter repos so you never begin from an empty folder. Copy-paste instructions for every tool. Ten to fifteen hours a week, and none of them spent on syntax.
Raise the first two, cut the last two, and the thing gets built. That’s the entire design brief for everything below.
What you get
Two books, thirteen templates, and every error decoded.
The Field Guide — 71 pages
The twenty mistakes that kill non-technical founders, each with exactly what to do instead. This is the part that changes how you think, and it takes 45 minutes.
The Workbook — 110 pages
Thirteen working documents in seven parts. Ruled fill-in lines, write-in boxes, and prompts built to be pasted straight into your AI.
The Idea Bank — 50 pre-scoped projects
Fifty real ideas with the one-sentence core already written and the complexity already scored. "I don't know what to build" is the most common reason people never start. Solved on page one.
The Error Atlas — 25 errors, decoded
What each one actually means in English, the usual cause, and the exact message to send. The difference between a two-hour spiral and a four-minute fix — and it's why people don't quit at the third failed attempt.
The Two-Brain Method
Run Claude chat as your Architect and Claude Code as your Engineer, with you in the middle deciding. The single highest-leverage structural choice I made, with the exact setup instructions and the one-line shortcut that automates it.
The Operator Loop
Context, Constraint, Build, Evidence, Correct. Every feature and every fix runs the same five steps. Once it's automatic you can build almost anything at this scale.
The Claude Code Setup Guide
Which plan, which model, which settings, and the one keyword that actually makes it think harder — plus the three that don't, which most of the internet gets wrong.
The Context Primers
Four copy-paste prompts that hand off everything you've decided to a fresh session, so you stop losing an afternoon's progress every time the window fills up.
The AI Cost Control Guide
How not to wake up to a $600 API bill. Nobody addresses this and everybody quietly worries about it.
Editable Word templates + plain text
The same content in three formats: read the workbook, type in the Word files, paste the plain text into Claude. Whichever suits what you're doing.
Get all of it, right now.
Forty-seven dollars. Less than one month of the AI subscription you’re already paying for, and roughly one three-hundredth of what an agency quotes to build the same thing badly.
Instant download
The 4 Week MVP Starter Pack
- The Field Guide — 71-page PDF
- The Workbook — 153-page fill-in PDF, 18 documents
- Editable Word templates + plain text for pasting into Claude
- The Claude Code Setup Guide — plans, models, settings, context primers
- The Idea Bank — 50 pre-scoped projects, scored
- The Error Atlas — 25 errors, decoded
- The Idea Viability Brief
- The MVP Specification template
- The Architect setup — copy-paste instructions
- The Master Build Request template
- The Debugging Packet
- The Deployment Checklist
- The AI Cost Control Guide
- How to Get Your First 100 Users
- User Outreach 1001 — scripts and Boolean searches
- SaaS Monetization Models
- The Do's and Don'ts of VC Meetings
- The Window of Diminishing Returns
- Cowork, Skills and Plugins
Secure checkout by Stripe. Delivered instantly to your email.
Zero risk, and I mean zero
Read It Or It's Free
Read it. If you don't find at least one thing that changes how you work with AI, reply to your receipt within 30 days and I'll refund you. Keep the book and the toolkit either way — I'm not going to chase you for files.
This is for you if
- You have an idea and no engineering background
- You've tried an AI coding tool and got stuck
- You want to build a real, small, deployed thing — not learn to code as a career
- You'd rather ship something imperfect this month than plan something perfect all year
This is not for you if
- You want a prompt pack and no work
- You're looking for a guarantee that a specific amount of money appears
- You're an experienced engineer — you'll find most of this obvious
- You aren't willing to spend ten hours a week for a month
Fair questions
Will this make me money?
I hope so — genuinely, I hope you build something that changes your life. But I'm not going to promise it, because I don't know your idea or your market. What I can hand you is the method for getting a real product deployed. The money is downstream of shipping something people want, and you can't ship what you never built.
Do I need to know how to code?
No. You need to be able to follow instructions, run a few commands, and read an error message. The book teaches the rest as you hit it, and the glossary covers every term you'll run into in the first month.
Is this just prompts?
No. There are templates, and they're good ones, but the value is the operating system around them: how to scope, specify, direct, test, and correct. Prompts without that are a slot machine.
Which tools does it use?
GitHub, Vercel, Supabase, Claude and Claude Code — because that's what I used. I'm not paid by any of them. The method works with other tools; the book says so explicitly.
How long does it take to read?
About 45 minutes for the Field Guide. The Workbook is a reference you use over four weeks, not something you read end to end.
What if I want more than a book?
The System ($297) is the full video curriculum with the reference application built end to end. The Lab ($2,500) is four weeks with me in it — and you get a deployed app or I keep working with you free until you do.
What if I hate it?
Read it. If you don't find at least one thing that changes how you work with AI, reply to your receipt within 30 days and I'll refund you. Keep the book and the toolkit either way — I'm not going to chase you for files.
Pick the date. Tell someone. Build the small version.
A year from now your idea will either exist or it won’t. That’s genuinely the whole decision, and it costs forty-seven dollars to be on the right side of it.
Get the whole system — $47Don’t be Bill.