Software for One

Software has always justified itself by scale. You build something because a million people will use it. The cost of writing code demanded that math. One engineer's time had to be worth more than one person's problem.

AI just broke that equation.

Right now, people are building software nobody else will ever use. Not prototypes. Not side projects waiting for users. Finished, working systems designed for an audience of one.

I'm one of them. I built a daily family newsletter today. It pulls in what needs to be shared for family logistics — only that — and emails my wife every morning with everything she needs to know. It took one sitting. Not because the technology is trivial, but because the cost of building it collapsed to nearly zero.

That newsletter will never have another user. It doesn't need one. It references our kids' schedules, our specific calendars, our household system. It's shaped entirely by the life it reflects.

What personal software looks like

Nobody asks "what's the best journal?" A journal is one of the most personal things you can make. What you write, how you write, why you write at all — it's all shaped by who you are. No two journals look alike, and they're not supposed to.

Personal software works the same way. My system is a set of markdown files, a few scripts, and an AI that reads them every time it wakes up. Someone else's might be a 3D explorable interface. Another person's might be AI helping manage a spreadsheet. There's no template. There's no correct architecture. The system reflects the person.

Software used to require enough effort that building for yourself was a luxury. You'd use the closest existing product and live with the gaps. The spreadsheet that almost tracks your budget. The calendar that almost coordinates your family. The note app that almost captures how you think.

AI closed the "almost."

It compounds

Each thing you build makes the next thing easier, because context accumulates.

A week ago I brainstormed a list of everything my family could use help with — schedules, finances, travel, household logistics. Just a voice note, thinking out loud. A week later, it's all built and working. The family newsletter took one sitting because the calendars were already connected, the contacts already mapped, the household responsibilities already structured. Every file I added became material the next thing could use.

Trip planning shows this best. A generic AI can plan a trip. Mine checks four family calendars for conflicts, pulls passport expiration dates, and flags school schedule overlaps. It knows which airlines we prefer, that one kid needs a window seat, and what our budget looks like. It knows we like walkable neighborhoods, that we prefer to take it easy and not cram too many things into a day, and that we've already been to certain cities.

Same model anyone can use. The difference is a folder of files — and each file makes every future query richer.

Why it matters

Personal software was never economical. The cost of building always demanded scale. AI collapsed that ratio, and now one person's problem — or one family's — is worth solving on its own.

No user research. No feature prioritization. No compromise between what you need and what the market will pay for. You just describe what you want, and it exists.

Your life is specific. Your software should be too.