Earlier this year, during lockdown, I watched Halt and Catch Fire, which had been recommended by Jason Kottke and Robin Sloan. It’s a bit uneven, but it captures the excitement of making something new and the frenzy and FOMO of trying to always catch the next thing over the horizon before it gets here.

A few years pass between Seasons 3 and 4, and when Season 4 opens we learn that one of the main characters has spent those years primarily in a basement, watching the Web grow from nothing. In his passion, and his unshakeable conviction that this is the start of something important, he writes down the websites he finds. He amasses hundreds of post-it notes with URLs. A catalog. A list. No taxonomy at all, no abstraction, just a raw list at the lowest possible level, as close to sense data as you can get.

Eventually he turns this into something more, but I think the idea here is undervalued. We come at the world with so many preconceptions, and it takes real work to just see what’s there. Gwern develops the idea of seeing through levels of abstraction, which is closely related, but he’s mainly talking about trying to achieve some goal in the context of an existing system. What I’m exploring here is more about gaining knowledge. The advice, then, is to begin by putting down what you see in the simplest possible terms and resist the temptation to categorize and theory-build until later.

This is a continual exertion. We love patterns because they allow us to think less, and we hate thinking because it’s so metabolically costly.

Of course, at every level, from the neural to the organizational, there is a constant back and forth between theory and data: maybe this is just an exhortation to stay in “data” longer than seems reasonable. Once you go to “theory” it’s hard to come back.

Edit Log:

  • 2020/06/22: fixed a typo. Also remembered that I own but haven’t read Umberto Eco’s Infinity of Lists, described here, which explores the same idea.

  • 2020/09/27: A related quote from Paul Halmos, which I found on this blog):

    “A good stock of examples, as large as possible, is indispensable for a thorough understanding of any concept, and when I want to learn something new, I make it my first job to build one.”