Skip to content
Sam Millan

← Writing · The starting point

First captured 16 April 2026 · revised 12 October 2026

The SOVEREIGN thesis

Own what compounds. Rent capability when it helps. Keep the memory, the judgement and the ability to leave.

Own the things that compound.

I keep coming back to the same question: how much of my life is actually mine?

My notes, the things I’ve learnt, the work I’ve made and the decisions I can stand behind. Those accumulate. I want them to remain useful when the software changes, a subscription ends or I choose a different way to work.

That is the idea behind SOVEREIGN. Own what compounds. Rent capability when it helps. Keep the memory, the judgement and the ability to leave.

Start with the memory.

In an April reflection I asked what would break if the company behind my AI changed its pricing. My answer was the conversational layer. The notes underneath should still be there, ready for another tool to read.

That distinction matters more to me than loyalty to a particular model. The model helps me work with what I know. The record of what I know should survive the model.

Owning files does not make every interaction private. If I give those files to a hosted model, that processing still happens elsewhere. Portable memory and local intelligence are different steps. I want to be clear about which one I’m taking.

Keep the first attempt.

There is another kind of dependence that survives even if I own all the hardware. I can ask a tool what I think before I’ve tried to think for myself.

I built Ponder around the opposite order. Get a question. Answer it myself. Then use the tool to challenge the answer, organise it or connect it to something I’ve said before.

Sometimes the useful response is to reject the question. Sometimes the useful work is the difficult first paragraph. I want help with thinking that leaves me better at thinking afterwards.

A life needs room outside the record.

I also want to notice when documenting something changes the thing itself. A run can become a post before I’ve finished running. A project can become attractive because it makes a better story.

Some things improve with an audience. Some need to stay private. I don’t have a clean answer for where that boundary sits, especially when privacy can also become a comfortable reason never to publish.

That tension belongs here. Ownership includes choosing what to share, and taking responsibility for the things I do decide to put my name on.

One idea, a wider horizon.

The wider project reaches into my body, home, data, mind and money. Those are places to apply the same question. They are not five claims that I’ve already solved my life.

The writing starts with data and mind: what I keep, what I outsource, what I practise and what I choose to make public. The essays explore different parts of that argument. Some develop it. Some expose a limit. Some end with a question I still need to answer.

The test stays practical: if the company behind this changes its mind tomorrow, what happens to me? Start with the answer that bothers you. Bring one thing back under your control.

Revised from my sovereignty thesis, April Ponder reflections, the dependence audit and the parts I don’t publish. This is a working thesis, with room to change as the evidence changes.