YOU MADE THE THING THAT RUNS YOU NOW

He wrote the first line of it at 5:41 on a Wednesday in a room nobody else entered.

The second bedroom of a rental in Cedar Rapids that served as an office because the apartment had no office and the project had no funding and the only resource available at 5:41 was time.

Infrastructure code.

The kind that runs beneath other people’s products the way plumbing runs beneath a building nobody photographs.

He wrote it before his shift at the data center and after his daughter went to sleep and on Saturday mornings when his wife took her to the park and the house went quiet in the way a house goes quiet when one person has stayed behind to work on something nobody asked for.

Seven years.

No users.

No loop.

Each module was better than the last because writing the previous one had taught him something the documentation did not contain.

The improvement was real.

The improvement was invisible because zero users times any quality multiplier is still zero.

He was not building toward a launch.

He was solving a problem that interested him more than the people who might eventually need the solution.

In February of the seventh year, an engineer at a company whose infrastructure failed every eleven days found his repository through a search query he had never optimized for.

She downloaded it.

Deployed it.

It held.

She told a colleague.

The colleague told a Slack channel.

The channel told a newsletter.

And the loop closed.

Each user who adopted it told others who had the same problem.

Each adopter who filed a bug drew three more who needed the fix.

Each contributor who improved a module attracted contributors who wanted to improve the one beside it.

The product was producing more of itself.

He did not push harder in February than he had in January or in October of the first year or on any of the twenty six hundred mornings he had sat in that room writing code nobody ran.

Nothing changed in what he was doing.

The concentration crossed a threshold and the feedback engaged and the mathematics did what mathematics does when the output feeds the input.

It ran.

By June the repository had more contributors than he could name.

By September it had more users than he could count.

He sat at the same desk in the same room watching the thing he had built begin building itself.

Each release attracted the developers who produced the next release.

Each integration generated the dependency that demanded the next integration.

The output was feeding the reaction now.

And the version of himself that had existed in that room for seven years was being dissolved by a version the loop required.

The maintainer.

The public face of a project that had outgrown its second bedroom.

The person answering issues at midnight because the loop does not sleep and the loop’s demand is a product of the loop’s growth and growth does not negotiate hours.

He had started the reaction.

The reaction was running him.

I ran a smaller version of the same machinery before I had a name for it.

Not code.

Sentences.

I wrote for four years in a silence so complete that the silence became the texture of the work.

Each piece taught me something the previous one did not contain.

The mechanism that converts a noun into weight.

The rhythm that holds a reader past the point where content alone would release them.

The sentence that earns the right to the sentence that follows it.

The improvement was real and the improvement was invisible because the loop had not closed.

No reader who shared produced more readers who shared.

No piece that landed generated demand for the next piece.

I was producing output that did not produce more output.

Subcritical.

The word I did not know then.

I called it not working.

I called it evidence that the sentences were not good enough or the audience did not exist or the timing was wrong.

Every explanation pointed at me.

The substrate.

Never at the loop.

Then one piece caught.

Not the best thing I had written.

Not the most carefully assembled.

One that happened to land where someone with reach would find it.

The reach produced more reach.

Within weeks the mathematics was visible.

Each reader who forwarded it produced readers who forwarded it and each reader who arrived found the previous work and the previous work fed the present work and the present work drew the next reader.

The loop closed and I was inside it and I could not tell where my effort ended and the loop’s momentum began.

The same morning.

The same desk.

The same hours.

A different dynamic that had nothing to do with what I was doing and everything to do with the concentration of product on the other side of the screen.

Here is what is underneath.

A process whose product is also its catalyst.

The output feeds back into the input.

The more product that exists, the faster new product forms.

This is not momentum.

Momentum is mass times velocity.

It continues in a straight line until friction stops it.

This is autocatalysis.

The product accelerates the reaction that produces the product.

The mathematics has three phases.

The lag.

When product concentration is low, the growth rate is low.

The machinery is running but the numbers are too small for the progress to register against the noise.

The reaction appears dormant.

It is not dormant.

It is accumulating below the threshold of human patience.

The explosion.

As product accumulates past a critical mass, the rate surges.

More product means more catalyst means more product.

The curve bends upward.

From outside it looks like magic.

From inside the equation nothing changed except the concentration.

The saturation.

Substrate is finite.

As the available material depletes, growth decelerates.

The curve flattens.

No autocatalytic process runs forever.

The thing that fueled the explosion is the same thing that limits it.

Three phases from one equation.

But the equation has a precondition that changes everything.

The activation barrier.

Autocatalysis does not start from zero.

The product must exist before the reaction that produces the product can begin.

Zero times any rate constant is zero.

The system sits in silence until a single event crosses the threshold.

Below the threshold, each chain fizzles.

Above it, each chain multiplies.

The transition is not a slope.

It is a cliff.

On one side you push and the thing you push does not push back.

On the other side the thing you pushed is pushing itself and you are inside it watching it accelerate on fuel you no longer need to provide.

The industry that sells instruction to people on the wrong side of the cliff has never named the cliff.

Consistency.

Post every day.

Ship every week.

Show up and the algorithm will find you.

Consistency provides the substrate.

But substrate is not the mechanism.

The mechanism is the activation barrier and nobody mentions the activation barrier because the activation barrier is not a behavior and behaviors are what the industry knows how to sell.

You can be consistent for seven years and never cross the threshold because crossing requires the product to reach a concentration where the feedback engages and concentration is a function of distribution and distribution is a function of existing reach and existing reach is the thing you do not have because you have not crossed the threshold yet.

Circular.

Exactly.

That is the bootstrap problem.

The system needs its own product to start producing its product.

Consistency does not solve it.

Consistency fills the container.

The threshold determines whether the container ignites.

Hustle culture.

Ten thousand dollar keynotes selling intensity to an audience standing on the wrong side of a cliff that intensity cannot cross.

Work harder.

Rise earlier.

Want it more.

A subcritical mass of fissile material produces chain reactions that fizzle regardless of the bombardment.

The geometry has to shift.

Enough product has to find more product.

The loop has to close.

No amount of substrate compensates for a loop that has not closed.

Grit.

The seven letter word that won a MacArthur grant and launched a thousand school programs.

Passion plus perseverance over time.

The equation is missing a term.

The feedback coefficient.

Passion plus perseverance in a system where the output does not catalyze more output produces a person who worked very hard for a very long time below the threshold of visibility and the culture cannot distinguish between noble and subcritical because it has no word for the difference.

Overnight success.

The story that begins at the explosion and amputates the lag.

She came out of nowhere.

She was in the lag phase for seven years.

It just took off.

It crossed the activation barrier and the mathematics did the rest.

The narrative credits the explosion to the person because the lag was invisible and the threshold is not a character in any story the culture knows how to tell.

The entire apparatus assumes the person is the mechanism.

The person is the substrate.

The mechanism is the loop.

And the loop does not care who provided the substrate.

It cares whether the product concentration crossed the line.

He is three years past the crossing now.

The repository has contributors he has never spoken to writing modules he did not design for systems he has never seen.

His name is on the project the way a founder’s name stays on a building long after the building generates its own electricity.

He maintains.

He does not build.

The building builds itself.

Each release attracts the developers who produce the next release.

Each integration generates the dependency that demands the next integration.

The product is the catalyst and the catalyst has no off switch and no field for the preferences of the person who once sat alone in a second bedroom in Cedar Rapids writing the first line of something that now writes its own lines through the hands of people he will never meet.

He works more hours now than he did in the lag.

Not because the work requires his effort.

Because the loop requires his presence and his presence has become a product of the loop and the loop does not distinguish between what he wants to build and what the product requires him to maintain.

His wife told me she noticed it in March.

He used to hum when he coded.

Low and almost inaudible.

The sound a person makes when the work is theirs and the only pressure on the keys is the pressure of the problem and the problem is enough.

She has not heard the humming since the loop took over.

He told me in August that he misses the room.

The quiet one.

The seven years when the code was subcritical and the feedback was zero and the work belonged to him because no external process was amplifying it or directing it or requiring that it continue in a shape the loop had already chosen.

The lag was the only phase where the work was his.

He did not know this while he was in it.

He could not have known because the knowing requires the contrast and the contrast requires the explosion and the explosion is the thing that consumed the ownership it was supposed to reward.

The crossing did not feel like arrival.

It felt like a door that opened in one direction.

He was through it before he noticed it did not open back.

I know the mechanism and I am inside it.

I know the lag is accumulation below the threshold of seeing.

I know the explosion is the mathematics of a feedback loop crossing a cliff that effort alone cannot cross.

I know the saturation is the substrate running out.

I know that none of these phases belong to the person.

They belong to the equation.

The person provides the substrate.

The equation decides the rest.

I still write each morning.

The feedback has engaged.

The product is producing more of itself and the production feels like mine the way his code felt like his in the seven years before the loop closed.

The feeling is less reliable now because the loop has its own rate constant and the rate constant does not carry a label distinguishing the part that is me from the part that is the mathematics.

I cannot find the seam.

I started the reaction.

The reaction is running.

I do not know which mornings belong to me and which mornings belong to the loop.

The knowing does not stop the reaction.

It stops one thing.

I no longer mistake the explosion for the beginning.

The beginning was the lag.

The lag was the only part that was entirely mine.

And I did not know its name while I was in it.

These are words. The mechanism they describe is not words. You will have to look for yourself.