On getting better every day
I'm trying to get better at programming every day. Even if it's just a little: most of the game is avoiding zero outputs, after all.
Here are some notes on the process:
- I've got a spreadsheet where, for every day, I give myself a binary grade and, if it's "yes, I got better," a quick note about how.
- Actually setting the goal of improving, and then trying--in however crude and coarse-grained a way--to hold oneself accountable, looks and feels very different from things I've tried before. It's just so easy to want to get better, decide to read a book, and read it in a way that accomplishes very little. I'm pleasantly surprised at how good (and challenging) it feels to have a mechanism to keep the goal in mind every day.
- It's not trivial to figure out how to get better, and it's easy to set goals that feel like progress but aren't. Many poker players ask what book they should be reading: I love poker books, but most of these players don't do nearly enough to translate their learning into actually improved play, and they wind up with zero benefit. I love broadening my knowledge, but I can't give myself credit for (e.g.) learning "hello world" in a new language.
- Happily, there is a familiar(-to-me) feeling that goes something like this: "I really should figure this out properly, but I'm in the middle of a task so will just do the magic incantation that happens to work." I'm training myself to recognize this feeling and put an item on my "actually getting better" list. Then, as I have dedicated getting-better time, I can refer to the list.
- These days, however, just doing my best not to get stuck in local maxima of generative AI competence is often the best way to get better.
- Getting better every day is probably not the optimum, at least not for me, right now. I get paid to get things done: sometimes that means that whole work days get spend exercising skill instead of building it.