What are the Best Hammers in the Rationalist Community?

by Bound_up1 min read24th Jan 20187 comments

17

TechniquesRationality
Frontpage

I've toyed with the idea of giving an intro to rationality, hoping I can find the 20% of the material that provides 80% of the gains (or the 5% that provides 50% of the gains).

Here and there, I've also been asked what the rationality community is about, and I've struggled to give an effective summary. I usually say the 3 keys are seeing in terms of concepts rather than words (a la rationalist Taboo or hyperdimensional thingspace), fighting motivated reasoning (since most people are already strong enough to prove others wrong and just need to turn that force upon themselves), and correcting for the heuristics and biases. But are these really the top 3 most important pieces? Even if they are, what other Hammers would take 4th, 5th, and 6th place?

The Hammers and Nails post got me thinking again: What are the most powerful pieces of rationalist lore? If I had 1 day to tutor a brilliant, but untrained mind, what are the most powerful techniques that can most easily be taught?

The comments under Hammers and Nails are full of people saying what their favorite and most powerful techniques are, but I'd like to make an explicit invitation here to share. What are your hammers? (see https://www.lesserwrong.com/posts/QzBuuNEqJGQFeWM4f/hammers-and-nails)

7 comments, sorted by Highlighting new comments since Today at 12:05 PM
New Comment

I cannot say much about CFAR techniques, but I'd nominate the following as candidates for LW "hammers":

Of course, the list is not exhaustive.

The most important LW idea for me was the Mysterious Answers to Mysterious Questions sequence. Someone can read it from start to end and come out a different person. Basically it teaches you how to get rid of "floating" beliefs that are underconstrained by data. My summary doesn't do it justice - Eliezer did an incredible job with that sequence, there's tons of metaphors and examples to make it stick. I reread it every couple years as a kind of spring cleaning.

This sequence was also quite profound to me. Maybe we should do a poll on the most powerful Sequence and reorganize them in that order.

I am very interested in any input on improving the structure and order of the sequences, though I also think it's a super hard problem and the correct solutions might require significant amounts of innovation and care.

Stating subjective probabilities of outcomes before the event happens.

Examples:

(1) If you ask for people whether they will come to your meetup, instead of asking the binary Yes/No, ask them for their probability.

(2) Elon Musk reports that he feels it's easier to make couragious decisions such as starting SpaceX and Tesla after he thought through the probabilities.

(3) Prediction-based medicine

(4) Determining odds for bets via stating your probabilities before. Based on the post Even Odds a member of our dojo created a tool that implements the math.

I've also been asked what the rationality community is about, and I've struggled to give an effective summary

Nitpick: the rationality community is not about anything simple, though the rallying flag is Rationality: A-Z. (cf 'The Rallying Flag is not the Tribe' by Scott)

It seems to me that in the context of Hammers and Nails, rationality is a community of Nails organized around the problems of heuristics and biases. The problem will be considered solved when we can reliably become Hammers because we have developed successful strategies that can be applied whenever such a heuristic or bias may reveal itself.

I think it is worth distinguishing this point because the heuristics and biases are innate to ourselves, so we can expect to have no choice but to become a Hammer in order to think clearly and correctly about whatever object-level problem we encounter.

For most powerful rationality technique I nominate shut up and multiply, because it avoids lots of heuristics and biases by deliberately invoking System 2 and it has a very strong track record (such as technology).