I’m a huge fan of the set of concepts named after Vilfredo Pareto. I know nearly nothing about the man himself, but a quick skim of his Wikipedia page suggests that he had some vague visions which were correct, but maybe didn’t quite nail the terminology/definitions himself. Still, it seems he was able to do enough to get us, I don’t know, maybe 80% of the way there.
Concepts:
- Pareto principle: aka. the 80/20 principle, that 20% of the effort can get you 80% of the desired outcome, and the remaining 20% will take 80% of the effort.
- It’s useful in two ways: when you can identify that you’re in a situation where 80% of the nominally desired outcome will suffice, you can save 80% of your time/effort!
- And when you are not in such a situation, budgeting time with the knowledge that the latter bits will take disproportionately long helps make more calibrated predictions.
- Pareto efficient: a situation is Pareto efficient if improving in any one desired dimension requires sacrifice in another desired dimension.
- Pareto frontier: the set of options which are Pareto efficient.
- We are often optimizing for multiple objectives in real life! Sure, I’d like to find the nicest apartment, but I’d also like for it to be as cheap as possible. Sure, I’d like to be maximally agentic, but I’d also like to be as conscientious as possible. Often, it’s easy to get stuck either trying to optimize in one dimension without realizing that you are already at the frontier, and further optimization will be costly, or to get stuck evaluating your local tradeoffs without realizing that you have a ways to go before you get to the actual frontier.