The Heuristic Heuristic

Wrong but useful generalization: 50% of rationality is about telling you what sort of strategies you can execute at valuable times. 30% of rationality is about telling when your normal strategies do something you probably didn't intend. The last 20% is about when your default strategies are actually still useful for some second-order reason you didn't consider.

A lot of the shiny, tasty stuff is in the first 50%. It's where techniques and fancy names get assigned to sequences of actions. Cognitive biases fill up the other 30%. The last 20% is where things get interesting; it's where you have to adjudicate which technique to use, if any, and what biases are relevant, if any.

If you spend most of your time learning about your existing heuristics as well as new ones, one of the most important skills to master is the "heuristic heruistic", that is, knowing when to apply advice, and when the law of opposite advice holds.

What does the skill of applying the right skill look like?

  1. Accuing enough experience to know, based on other factors
  2. Being open to the other side (ala truthseeking via the Double Crux attitude)
  3. Being good at executing on strategies
  4. Understanding the underlying dynamics at play
  5. Probably some sort of FDT-esque reasoning that understands what level of generality actually generalies

Last Updated: 2020-07-12 22:36
First Published: 2020-07-12 22:36