DRMacIver's Notebook

You should try bad things

You should try bad things

Earlier Avy asked me the following question:

Why would you read something if you expect to be terrible?

My answer started with a few specific reasons, but finished with the following general observation:

Also fundamentally... If you never to anything you think is going to fail, you'll repeatedly fall victim to confirmation bias and never leave your comfort zone. Sometimes it's worth branching out just to see what happens when you do

This is a problem I talked about in There's no single error rate: If you only only observe the consequences of one part of your selection process (what happens when you try things you expect to be good) you've given the other part (what happens when you try things you expect to be bad) free reign to be basically arbitrarily wrong. This means you're missing out.

In particular it means you're missing out systematically, because you will never correct on traits that you think of as signalling something terrible because you will always avoid them.

I do a lot of reading outside of my comfort zone, because that's where things I don't know about live. Sometimes this requires gritting my teeth and reading something bad because it contains information I will get nowhere else (I find a lot of therapy reading like this - it's written in a style I really dislike), and sometimes this requires reading something I expect to be bad and finding out that it's actually very good.

If you only try things that you expect to be good, your interests will narrow over time, because you will learn that more things that you thought were good are actually bad, but you will never learn that things you thought were bad are actually good.

If you're certain something's going to be bad because you've tried it enough times to know, skip it sure. If something is very high effort, or very high risk, you similarly shouldn't push yourself into doing it. But if you can find or create relatively cheap and safe opportunities to try out something that might be bad, try it out. You might learn something important.