Goals.

Written down and put somewhere public, which is the only version of a goal that has ever worked for me. Two lists, same rule for both: something another person could look at and tell whether it happened. The first one keeps its own score.

Be brave. Nothing on this page was a safe bet the day I wrote it down, and that is the whole point. A goal you are certain of is not a goal, it is a schedule. Taking the risk is the part that costs something.

Work

• A billion downloads a month The whole point of the tooling, stated as a number someone else can check. Today it is 17.4M a month across 421 published packages. That is 1.74% of the way, or about 57× from here. It resyncs daily, so this page marks its own homework. 17.4M of 1B a month
• Stacks v1.0 The framework stops being a moving target. A stable API, documentation that matches it, fast and energy efficient by default with published benchmarks to show it, and a version number I am willing to ask other people to depend on.
• Wildloop Ship the territory game properly: captures the server can stand behind, clubs and challenges people come back to, and route records worth chasing.
• HQ.training Make it the thing coaches reach for first: training analysis that earns its place next to the billing and the scheduling, rather than sitting in a separate tab.

Outside

• Pacific Crest Trail Northbound, 2,650 miles, aiming to average 60 a day. I do not know yet if that pace holds. Finding out is half the point.
• Race it Once the trail is done, start pinning on a number at mountain races to see where I stand.
• Train in the open Run the build in public the same way the code is: the days it goes well and the days it does not.
• Every national park All 63 in the US. I have been to 20 so far. 20 of 63 parks

The training behind the running half is public: two years of it, on HQ.training, pointed at my own weeks.