I have a heuristic that you basically shouldn’t try to optimize for multiple things at once. This is commonly known for charitable donations, but applies in many other cases as well: pretty much whenever you try to optimize for two things with one action, you’re cutting down the search space by a huge amount, so it’s probably not worth it.
Examples:
When choosing a job, don’t try to find something that’s both lucrative and does social good. You’ll end up in something like public interest law or tech work for nonprofits. This pays vastly less than the highest-paying jobs, like finance, and does vastly less good than the best direct work jobs, like working for an effective altruist organization, to the extent that either extreme is most likely better.
If you’re starting a company for altruistic reasons, you should either focus entirely on making as much money as possible and donating it, or entirely on creating as much value as possible through the startup. For instance, if you’re maximizing income you should be stingy with equity, whereas if you’re maximizing value-creation you should be generous. Also, for income you should definitely start your business in the US, but for value-creation you may want to consider a developing country instead.
I think this also holds in many other domains, e.g. hobbies, classes, and basically anything else where you choose one or more things based on multiple criteria. Thoughts?
Comments
This works in cases where you can do multiple things (choosing different charities that fill different roles), but your second two examples only work because there’s a pretty good exchange rate between money and value in the form of good charities. This heuristic won’t work well in cases where you can only do one thing and want multiple things out of it. If you’re buying a car you don’t want to optimize for only one of fuel efficiency, internal volume, speed, or price.
OK, my brain has two semi-competing heuristics here. On the one hand: “things vary on many axes but there’s usually only one important axis”; on the other, “extreme values on one axis tend to be costly along others”.
Competitive scenarios where most people agree on the axes tend to cause the latter heuristic to apply more. In Jeff’s car example, for instance, all the car companies are competing for your dollars, and they make different tradeoffs along certain axes in order to appeal to different types of people.
On the other hand, Ben’s job example is an instance where the first heuristic applies: not very many people are optimizing for purely lucrative jobs – in other words, the axis you care about is rarely the one that others care about – but money converts well into a lot of things, so you can select heavily along that axis and end up doing pretty damn well.
The corollary to these two heuristics is “find important hidden axes”. If you’re optimizing for something nobody else is caring about but is actually important, you can do well.