Very Very Fucking Simple

I was on holiday with my wife and, being a slight workaholic, completely failed at switching off.

Very relaxing. Very present. Very ordinary man of me.

My brain kept coming back to this learning application.

For a while, the premise had become a bit too clever. A bit too pure. A bit too focused on solving memory problems. A bit too anti-positioned against what already works in the market.

There was probably some desire in there to be different.

Not better. Different.

That is a suspicious move.

And the thing I kept coming back to was:

Keep it simple.

Not just simple.

Very very fucking simple.

A business needs a problem. Research helps you understand the problem. Then you make a solution. Then you position the brand around that solution. Then you build a first version based on assumptions and see what reality does to it.

If I simplified it even more, the whole thing became three moves:

Build a fantastic product for 20 to 50 users.

Validate activation, retention, revenue, and referral.

Then scale the content, the product, and customer acquisition.

That is almost annoyingly clear.

And because it is annoyingly clear, my brain immediately wanted to make it complicated again.

What about finance?

What about people?

What about operations?

What about raising capital?

What about all the backend machinery of a company?

Valid questions. Not fake questions.

But maybe some of them are later questions.

Maybe the first job is not to build the company-machine. Maybe the first job is to build the customer-machine.

The front end.

The thing the customer sees. The product they receive. The brand they perceive. The way they hear about it. The way they try it. The way they get value. The way they tell someone else.

Everything else should serve that.

Finance, operations, hiring, systems, capital. They matter. But early on, maybe they should not be allowed to become impressive hiding places.

Because it is very easy to feel like business is a complicated theatre.

A lot of meetings. A lot of status. A lot of structure. A lot of people making sure other people do work so other systems can keep moving so the customer eventually gets something.

And maybe that was needed more before.

But AI changes the shape a bit.

Not magically. Not “one person can do everything forever.” That is too clean and probably stupid.

But one capable person, or a very small aligned team, can now execute a ridiculous amount compared to before. Product, research, content, design, workflows, testing, support, analysis, documentation, automation. The leverage is not normal.

So maybe there is a new question:

How much company do you actually need before the customer machine works?

Maybe less than my old assumptions think.

That feels exciting, but also uncomfortable. Because if the plan is simple, there is less to hide behind.

You cannot keep polishing the architecture of the business.

You have to serve the users.

You have to make something useful.

You have to charge money.

You have to see if people come back.

You have to see if they tell someone.

Problems, solution, charge money, make money.

Obviously that is too crude. Also, maybe not crude enough.

The bias here is obvious. I am probably overcorrecting. Operations matter. Finance matters. People matter. Capital can matter. The boring adult stuff does not disappear because I had a thought on holiday.

But as a starting frame, I like the violence of it.

Build something excellent for a small number of people.

Make the loop work.

Then scale the shit that is already working.

It seems too simple.

But maybe that is the point.

Maybe the ego loves complexity because complexity feels like depth. Maybe simplicity feels suspicious because it leaves you alone with the actual work.

And the actual work, for now, might be very boring.

Find the problem.

Build the thing.

Serve 20 to 50 people insanely well.

See what happens.

Very very fucking simple.

Annoying, really.