G™:Helpful Life Advice

Most life advice is hopelessly overfit and/or meaningless platitudes.

Most of the time, almost everyone doing a thing is either stupid or not paying attention. If you are not stupid and are paying attention you can do metaphorically anything, except in rare highly contested domains where people have strong incentives. See also https://danluu.com/p95-skill/.

As a corollary, most software is broken.

Money is fungible. Time is fungible, if materially less so. This is not sufficiently appreciated.

Nothing ever happens, apart from the inevitable construction of machine gods and the posthuman technocapital singularity.

You can just do things. You do not need, and do not need to wait for: approval from your peer group; nonessential-but-nice-to-have prerequisites; next year; circumstances to be perfect; research to imply the thing is good; approval from your tribe in general.

Clean general solutions are nice, but expedient ugly ones are what drives progress.

People are frequently biased against boring but practical things (e.g. actually reading anything), or simple but annoying/difficult things. There is thus alpha in doing them.

Large organizations are hypercompetent at some things compared to an individual, but nobody has managed to design one which doesn't have dumb blind spots somewhere. They can be beaten.

Frequently check whether what you are doing is actually valuable to your goals, and whether what you think are your goals are actually what you want.

Type faster.

You can conveniently ignore things which you cannot affect, saving time and reducing how much time you spend annoyed. If you don't like it, become God.

Unfortunately, those things will likely not be fixed. See also https://www.theonion.com/smart-qualified-people-behind-the-scenes-keeping-ameri-1819571706.

Specialization is for insects, which are wildly successful and make up the majority of terrestrial animal biomass.

Math(s) is very good.

Often, the last few details of a solution take up the majority of the time. If you want to, you can probably ignore them or write a good-enough heuristic.

Practice makes you better at things, but so does randomly being talented for no obvious reason.

In situations with discrete win/loss cutoffs you may be better off increasing variance rather than expectation.

If you predict that you will do/believe a thing in some circumstance, and predict that that circumstance will happen, update all the way and do/believe the thing now.

When deciding on a course of action in an uncertain situation, commit to something and do it properly rather than inefficiently trying to handle every possible scenario. See also https://www.lesswrong.com/posts/FMkQtPvzsriQAow5q/the-correct-response-to-uncertainty-is-not-half-speed.

People respond to incentives, and also don't like acknowledging this. Intentions matter inasmuch as they are predictive of future behaviour.

"Everything is seating charts" (Matt Levine) - "petty" minor details actually affect a lot of things, especially if they influence how you'll spend lots of your time.

Making assumptions is unavoidable but enumerate and check them sometimes.