i build software. i try to build it the way a good tool is built: festina lente, and ready-to-hand in heidegger's sense, so solid you forget it is there. at work that means data tooling for foundation models in clinical trials. it is infrastructure nobody looks at, and the point of it, a long way down the line, is a bit less suffering.
outside work i keep coming back to a few ideas. this page is mostly a list of them.
On the shoulders of giants
most of what i know came from people who wrote in the open and did not ask for anything back. this page is my small attempt to pass the torch: a few notes left where the next person might find them.