Twenty weeks. Sixty posts. Never missed one. On Friday I missed one.
Nothing dramatic happened. I finished a twelve-hour shift, sat down, opened file, looked at it, and closed my pc.
I want to be straight about this because whole point of this series is that I do not edit the bad weeks out. I do not write these on a production line.
There is no buffer, no queue of drafts waiting. Each post gets built two or three days before it goes out, from that week's actual work.
That is deliberate, because a post written in advance is a post about a week that had not happened yet. Cost of that is simple: if the week takes everything, there is nothing left to write with.
Last week took everything. Seventy hours behind the wheel, seven days, and a weekend before it where I cleared an entire release so the app would not need me. Friday arrived and the tank was empty.
So the streak broke, and I would rather tell you that than backdate something.
The game
This week is lighter on purpose. And it belongs to the game. Scaling Laws is fifteen days old. In those fifteen days:
- 126 commits.
- Around 70,000 lines of C#.
- 672 automated tests across 64 files.
- 35 save-file versions, because the data model has changed under a live game thirty-five times and I will not break somebody's campaign to add a feature.
It is also, quietly, playable now. Not finished. Playable.
You run a frontier AI lab from 2022. You rent compute you cannot afford yet. You pick a model size you can actually supervise rather than one you can afford. You decide whether three weeks of safety evaluation is worth delaying a release, and a regulator will later judge you on what you had switched on day you shipped, not the day they arrived. Thirteen rival labs ship alongside you.
This week
- First test build. Ugly, honest, playable end to end.
- A demo after it, if the build says the economy is actually fun.
- Hiring screen polish. Three channels, and every candidate has a hidden reservation wage you have to negotiate against without seeing it.
And the taxi keeps running underneath all of it, because that is what funds the move.
Twenty weeks of not missing was a nice number. Turns out the number was never the point. Showing up on Monday after missing Friday is.
What is your streak, and what finally broke it?