The 5-minute rule (and why I quit chasing 10x)

EDITION 001 / SUNDAY APRIL 26, 2026

[ solo stack ]

Run lean. Stack smart. Compound forever.

For Christmas in middle school, I asked for Windows for Dummies books.

Not the toys other kids wanted. The books that would teach me what nobody in my rural town could.

By the time I was sixteen I'd bought a CD burner with my paycheque, installed it myself, and was selling pirated CDs to my classmates for twenty bucks a pop. Twenty-six years later, I'm still figuring out how to make technology pay rent.

The math has just gotten more interesting.

I built a fifty-person company from scratch. Did the MBA. Played the scaling game. Ran the place for years before I admitted something to myself out loud:

The conflict resolution drained me. The hiring and firing cycle drained me. The constant emotional weight of being responsible for fifty people's livelihoods drained me. None of those things were what I was good at. They were what the job required and what I had been trained to want.

Admitting it felt like a relief.

That is not a sentence my MBA professors would approve of. The whole framework I was taught (scale the team, scale the revenue, scale the title) assumes you want to be a manager. Mine never asked whether I should be one.

Six years ago I started doing something different.

When someone retired or moved on, I stopped immediately backfilling the role. Instead I sat with the question: what part of this job actually needs a human?

Most of the time, parts of it didn't.

The first big win was document automation. We were producing 300+ page custom documents, every single one tailored to a specific client. Each one took two weeks to build by hand.

I spent a year coding and building out the process. A year. That's not a sprint and it's not a hack. It was a slow, careful build with a lot of broken Saturdays.

Then it shipped.

Same quality. Same level of customization. The system has been running for six years and only needs minor tweaks once or twice a year.

That single project paid for itself a hundred times over. More importantly, it changed what was profitable. Work we couldn't take on at the old speed suddenly became revenue lines.

I didn't fire anyone. I kept the people I love working with. Some of them have been with me for over a decade. They're older. They don't want to learn this stuff on their own. So I build the systems. They run them. Everyone gets to do better work.

That is the part most AI newsletters get wrong. They talk about replacing people. I'm talking about keeping the people you actually want to work with, and removing the parts of their jobs that nobody enjoyed in the first place.

Here is the rule I've been running for six years.

Most operators chase the 10x wins. The big AI projects. The dramatic teardowns of entire workflows. I don't. I chase the 5-minute saves because they compound, and the math is wild once you actually run it:

The reason this works is psychological, not just mathematical. The 10x project is intimidating. You don't start it because you don't have a free Saturday. The 5-minute save is shippable in an afternoon. You start it because you can.

And then you stack them.

The 300-page document automation took me a year to build. The 5-minute saves take an afternoon each. I have something like fifteen of those running across my businesses now, every one of them quietly saving a small slice of time. None of them, individually, would be impressive enough to write a case study about. Together, they bought me back something I didn't know was on the menu: a life that runs on its own while I work on the next thing.

This year that includes places I didn't think I'd be able to work from.

I closed Q1 from Antarctica. As long as the ship had Wi-Fi, my businesses ran. The team handled what they handle. The systems I built handled the rest. I answered three emails in two weeks and nothing fell over.

That is what compounded automation buys you.

So. This newsletter.

I run multiple businesses. I lead small teams I've kept for over a decade. I spend most of my days building automations instead of managing people. I've been doing some version of this for twenty-six years and the AI tooling of the last three has changed what's possible more than the previous twenty combined.

For the next chapter, I'm writing it down publicly. Tuesday mornings.

Real numbers, real tools, real costs. What I'm building. What's working. What's broken. The stack I'm using and what I'm paying for it. The 5-minute saves I'm shipping that week and what they're worth in the long run.

The compound game, played one small win at a time.

[ THIS WEEK'S ASK ]

Hit reply and tell me one process you'd automate if you knew it would save 5 minutes a day. I read every reply, and the answers shape what I write next.

If this one hit, forward it to one operator who needs to hear it. That's how this newsletter grows.

[ BUILD LOG ]

[ STACK MATH ]

One number this week:

$0.34

What one run of my inbox triage automation costs me on Claude Haiku 4.5.

It saves roughly 4 hours of email work per week. Math: $0.34 × 5 runs/week × 52 weeks = $88/year. Buying back 200+ hours of my time.

That is the entire newsletter in one line.

[ NOTE ]

Writing this on a Sunday night, 42 years old, a glass of red wine in, the brackets of my brand-new logo glowing on the screen beside me. Twenty-six years in and still figuring out what I want to be when I grow up. Maybe that's the point.

See you Tuesday.

— [ solo ]