DRMacIver's Notebook
Studying specifics
Studying specifics
There’s something I deeply believe, and only very erratically manage to follow through on, which is that there’s a huge amount to be learned by studying something more than it seemingly warrants. This is in some sense the core of the post Overthinking overthinking which is what created the “Overthinking Everything” name that I now use for my newsletter and community.
One core reason for this is existence proofs. If you put more effort into solving a problem than it really warrants, you’ve now demonstrated that that problem was solvable, and you can potentially use that solution in other cases, gradually reducing its effort.
I think this relates to one of my points in Heavy up front learning requirements, which is that it’s worth spending some time heavily studying skills you use regularly, even ones that don’t seem like they need it. Sometimes just by really levelling up one specific task, you should be able to learn generalisable lessons.
The problem is of course… how do you actually do this? How do you pick what specifics to study?
One part of the answer is to just do which ones bring you joy, but I think there’s a key problem with this: The specific prescription is to study something more than it warrants. If it brings you joy to study it, hopefully you’re already using that joy to prompt you to spend more time on it.Right? Right?? Um, OK, this theory might have some problems.
You could study everything more than it warrants, but then you rapidly hit resource constraings. Where are you going to find time for that?
I think this is, often, the bottleneck on the joy version too: There are many things that if you just followed your curiosity you’d find would bring you joy, but there are too many things, and as a result you do none of them.
Instead, a better solution I think is to pick arbitrarily and capriciously. Perhaps, ideally, at random from those that you think seem like they would be interesting to do.
One of my recurring fantasy scenarios is what I’d do with an improbably large amount of money, and one of the regular projects I think about is just randomly sampling a few hundred people from the population and giving them near unlimited healthcare - extensive investigations of anything that’s a problem for them, suitable interventions (erring on the side of cheap and low risk ones), and then studying the data from that extensively to learn useable lessons for the broader population.
This strikes me as the right sort of general tool: By sampling randomly from the set of places you could intervene, you learn plenty of things that will generalise out of that.
I don’t yet have a very good idea of how to implement this though.