hayward@click:~/~/rants/2026-10-19-the-default-is-the-decision

what these numbers are

Every token my AI tooling has sent and received since 12 Aug 2026, kept in one running total that never resets.

is what goes in, is what comes back out. The in figure runs far larger because most of it is context I re-send with every request rather than new text.

Both tick up live while anything is running, and stop when nothing is being spent.

Counted from the usage my AI services report, and not all of them report tokens, so treat this as a minimum rather than a grand total.

The default is the decision

19 October 2026

I opened something I built a while back and could not tell myself why it was put together the way it was.

Not whether it worked. It worked. The parts sat in the right order, the names made sense, and the thing did what it was asked to do. What I could not find was the reason. Nothing in it was there because I had decided it should be. It was there because that is how the tool does it, and I had said yes at the time, and the yes had been so easy that I never noticed making the choice.

A build used to be made of those choices. Not the grand design, the one I argue with myself about. The small things. Where the settings live. What a thing is called. Which way round to do it, picked because I had been bitten by the other way once. Every one of those was mine, and I could tell you why, and the explaining is where I learned the thing in the first place.

Now the sensible answer arrives before I have a reason of my own. It is usually better than the one I would have reached for. That is the awkward part. A bad default is easy to spot and easy to refuse. The one that quietly beats me is the one I take without looking, and after enough of those I am running something that works and belongs to nobody.

I am not going back to the long way round to make myself feel useful. The long way was never a virtue, it was the price of not having anything better.

What I do instead is smaller than that. On each thing I build I pick a small decision and refuse the default, on purpose, even when the default is right, and I have to be able to say out loud why mine is right as well. Sometimes it is not, and I put it back, and that is the part that teaches me anything. I write the reason down next to the work, in my own words, so the next time I open the thing there is at least one line in it that is there because of me.

Everything still gets built. That was never the worry. The worry is getting very good at saying yes to answers I did not arrive at, and finding a finished, tidy thing on the machine one day that works perfectly and has none of me in it.

All rants · The cost used to kill the bad ideas