These Are Not Websites and AI Can't Save Us

This week I shipped four things:

OpenDev.studio — contracts that remember what happened. HavenHaus901.com — a venue system for people creating a scene. SwampSoulProductions.com — music production gear, people, and bookings in one place. LadiesPrimera901.com — a cleaning business with rails under the work.

They look like websites.

They are not websites.

Or maybe they are, in the way a placemat is a placemat. A thing on a table. A surface. A little rectangle of color under the plate.

But that is not really what it is.

The thing itself is never the thing itself when people are gathered around it, needing to be fed, needing to be remembered, needing some kind of order before the food arrives and everyone starts talking at once.

When I was up to my neck running businesses, I had a mantra I'd share with anyone willing to listen.

Solutions, not problems.

Which is funny, because business is mostly problems. The best businesses solve the biggest problems for the most people at the smallest cost. That's the clean version.

The real version is less clean.

The real version is people calling and texting and needing answers to questions you've answered before. Employees waiting on instructions. Customers waiting on callbacks. Vendors waiting on payment. Calendars half-right. Invoices half-sent. Promises made in one thread, remembered in another, and eventually disputed by someone who was probably tired when they made them.

So I grew to love the bomb.

Not the chaos exactly, but the moment when the chaos becomes visible enough that the right people can finally see it together.

My other mantra was best idea wins.

That never meant the loudest person wins, or the owner wins, or the lawyer wins, or the person with the spreadsheet wins. It meant the problem has to be cast in the right light, and then everyone around the table has to be willing to lose track of who said what.

A good team gets absorbed in the thing itself.

Someone says something. Someone else makes it better. Someone else sees the missing piece. The source disappears. The idea keeps moving.

Probably not unlike a writing room where a song is born and grows and finishes and no one really remembers who first found the line. Maybe the guy in the corner said the little note that made all the difference. Maybe that's why they share credit.

The past few weeks, I got to work with friends who share that instinct.

Each of them had already built something real. Not a pitch. Not a concept. Not a “brand.” Real businesses, with blood and sweat and late nights and customers and messages and staff and all the little human dependencies that make a thing work until it doesn't.

I had the luxury, or maybe the burden, of having seen that same movie in a dozen fields. Law. Bars. Construction. Rentals. Restaurants. Studios. Venues. Deals that were supposed to be simple and never were.

And then I went deep, bro.

Real deep.

Databases. System architecture. JSON files. Weird little objects with names and IDs and statuses and timestamps. A world that, for all practical purposes, was inaccessible to me a few months ago.

And I brought to the table a gang of very patient, very strange engineers in the form of giant language models that have forgotten more than I will ever know.

But AI can't save us.

That's the part I keep coming back to.

AI can write and summarize and suggest and classify and draw and argue and remember. It can move unbelievably fast. It can make one person feel like a small department.

But it cannot care about the problem.

It cannot know why the thing matters.

It cannot sit with your friend who has been carrying the business in her phone for three years and understand that what she needs is not another app, but time. A little dignity. A system that stops asking her to remember everything.

Most businesses do not fail because no one had an idea.

They fail because the idea keeps asking the same tired person to remember everything.

By the time I talked to Dustin, I had gone so far down the rabbit hole that it took half a day to ship something for Swamp Soul Productions that should have taken a small software team and a month of meetings.

But the problem was not software.

The problem was that expensive music production gear and skilled people were still living inside the old economy of who happens to know who, who remembers who has what, and who happens to pick up the phone when someone needs something next week.

The website is the surface.

Underneath it is a harness for gear, people, bookings, questions, and trust.

Then there was Sarah, who had built a cleaning business the hard way. Homes and clients and ladies who needed work and customers who needed help and families depending on all of it. At the end of the day, there wasn't much time left for the intentions. Bookkeeping. Marketing. More workers. More homes. A real system.

So Ladies Primera became less about a website than a way to catch the work before it disappeared into text messages and memory.

A place for the job, the worker, the client, the schedule, the payment, the next thing.

A little operating rail for a business that was already moving.

Ben and J started hosting shows during the pandemic and letting bands crash on their couches after the show.

Four years later, on their third venue, the DMs are mounting up. Flyers need to be drafted. Supplies need to be bought. Shows need to be hyped. Bands need somewhere to sleep other than the couch.

To jam econo is one thing.

To run a business like that is another.

Who has time to count the beans when you're creating a scene?

The emails stack up. The beer runs out. Someone needs the door count. Someone needs to know when load-in starts. Someone needs to be paid. Someone needs to be safe at 3 a.m.

Haven Haus is not a website.

It is a little frame around the beautiful problem of people gathering in a room to make noise together.

And then there are the lawyers.

God help us.

Tasked with turning words into a crystal ball. Asked to protect clients ten years out while everyone in the room is shouting, hurry up, get it done, we need to move.

Make the contracts mean something.

Make them hold the weight.

A thousand moving parts. A million decisions. Money, time, risk, approvals, notices, memories, all scattered across emails and meetings and PDFs and people who will eventually remember things differently.

OpenDev is the one I cannot stop thinking about because it is the oldest idea dressed in the newest clothes.

The contract should not be a dead thing.

It should remember.

It should know what kind of event just happened. A forecast. A fact. A request. An approval. A formal change. It should know who said it, when they said it, what authority they had, what gate it needed to pass through, and whether it changed anything at all.

Classic contract law was always trying to do this.

Give people certainty around time, money, and risk.

The technology is just what finally lets the record keep up with the project.

That's what I mean when I say these are not websites.

The website is the visible edge.

Behind it is the intake, the routing, the calendar, the payment link, the customer record, the worker assignment, the contract logic, the ledger, the next thing that needs to happen.

The old promise of software was efficiency.

The better promise is memory.

Not memory like nostalgia. Not memory like a scrapbook.

Memory like a rail.

Memory like a witness.

Memory like a quiet little object that says: this happened, at this time, from this person, with this effect, and now the next thing is this.

You can tinker on these websites and that's fine. I will too. The button could move. The copy could be better. The photo could be sharper. The logo could sit a little higher.

But the surface is not the experiment.

The experiment is underneath.

Math and hard code. Decisional logic. Fibered connections to other businesses already doing their own part of the work.

Little harnesses for living systems.

Snow globes waiting for the next time someone turns them upside down.

And they will be turned upside down.

That part is guaranteed.

The message will come in wrong. The job will move. The band will cancel. The customer will forget what she said. The contractor will remember it differently. The lawyer will find the gap. The dog will run into the woods. The boat will hit the wake. The child will grow up before you are ready.

AI can't save us from any of that.

Maybe nothing can.

But we can build things that remember where the pieces were.

We can build things that help people see the same problem at the same time.

We can build things that make the next right action a little easier to find.

And maybe that is the solution.

Not the website.

Not the model.

Not the magic.

The harness.

The thing that holds just long enough for the people doing the work to keep going.