Behaving as if you were trying to succeed

In a flat I shared a while back, for some reason we owned a wooden meat tenderising mallet (Disgusting object. Can't recommend). One of my flatmates was the main person who used it, so I assume it was hers, but it lived with all of the other kitchen utensils.

One day I discovered it had been put away still covered in bits of raw chicken. I politely expressed some disapproval of this fact. She, in turn, seemed vaguely baffled by this objection.

I don't remember her exact words, but it was something to the tune of "I did wash it! But it's hard to clean."¹

I, in turn, expressed the sentiment that if something is hard to clean then that means you clean it more thoroughly, or don't use it, but that putting things away with raw meat stuck to them was not among the acceptable set of outcomes.

She did rewash it, but I'm not sure my point really stuck.²

I've thought about this example over and over again since then, in various shared living situations, because I've run into a surprising number of people who are bad at washing up in a particular way: It's not that they're incompetent at the basic mechanical actions of washing up³, but that they don't check if the thing they've washed up is clean afterwards, and as a result a substantial fraction of things they wash up still have food stuck to them. It's usually not as bad as the raw chicken incident, but it's bad enough that I don't consider the things they washed up to be clean enough to put away without double checking.

As a result, I typically ended up rewashing stuff they've washed,⁴ which means they've put in a bunch of work for literally zero effect, and made me more annoyed at them than I would be if they hadn't done anything at all.

Similarly, back when I used to work in more end-user-facing software with an actual web interface and such,⁵ I used to deploy this really exciting magical code review technique that reliably found problems with people's pull requests: I checked out the branch, launched the web application, and tried actually using the feature. Often it had obvious problems.

This has become something of a bugbear of mine in general: People doing things as if the thing itself was the point, rather than a tool you are using to achieve a specific outcome, and not thinking about whether their actions will actually in any way help them achieve their goal. The result of this is often that they put in about 90% of the effort required to do it properly, and get 0% of the actual benefit because what they did didn't actually work.

The easiest way to fix this is to check whether what you did worked after doing it. Initially, this will slow you down a bit and add delays, but as you get good at it you will basically internalise the process and it doesn't slow you down much.⁶

I do, however, suspect that in many of these cases the actual thing you need to do to fix this is to actually care about the end result rather than just wanting to avoid being blamed for not doing the thing and hoping nobody will notice.⁷ You need to accept that actions you take have actual impact in the world and are not just something you are doing for the sake of it.

This is also important because there are many examples where you can't really check if it worked.

For example, I still semi-often see people wearing face masks in public. I rarely do these days, but consider this a completely reasonable choice. Unless, that is, you're wearing your facemask pulled down so it doesn't cover your nose. In which case what on earth are you doing? How have you made it to 2025 without looking up basic mask usage?

But you're never going to figure out that you did this wrong by checking. What are you going to check? Did I get a virus Y/N? That's an incredibly noisy signal, and tracing it back to the specific error you made is going to be almost impossible. The only way to get this right is to actually care about getting things right and then take whatever action is appropriate for achieving this.

In contrast, throughout the entire pandemic it felt like people were ignoring this. The reason people wear masks badly was reflected throughout their behaviour in the entire pandemic: They treated the point as adherence to the rule, and the pandemic as a social phenomenon, rather than thinking about actions as existing in a real physical world containing real things such as viruses that your actions interact with in meaningful ways. The mask acts, not as a tool, but as a talisman.

If you want to do things that actually work, and to avoid a whole bunch of wasted effort spent on things that don’t work, you need to let go of this talisman-type thinking and treat your actions as things with actual practical content. You need to care about whether they work, and you need to pay attention to the world and learn from it when they don’t.

Even with the best of intentions, you won’t always succeed. Certainly I don’t. Sometimes you’ll rush and forget to check something. Sometimes you’ll check something insufficiently carefully. Sometimes you’ll check something and misinterpret it, or otherwise make a stupid error. Sometimes you just won’t have the key piece of information you needed and won’t find it out until it’s too late.

This is fine. An attitude of caring to succeed is never going to get you to perfection, and I'm not asking you or anyone else to be perfect, but I am asking you to do better, and when you are failing, try to figure out why and fix that.

---

This post is a fairly lightly edited version of something I previously posted on my notebook that I decided was good enough that I wanted to promote to newsletter. If you’re feeling starved of my writing (sorry I’ve not been writing anything here recently), go check out the notebook, where I’ve been writing almost every day recently. It’s of highly variable quality and about a much more random collection of topics than here.

1. As a side note, I should observe that this was a long time ago. I no longer live with her but we're still in touch and she has become a vastly cleaner person in more recent years. We recently had a phone call in which I mentioned this story and that I’d written a notebook post about it and she said something along the lines of “oh god, I was the worst. I can’t believe none of you murdered me when we lived together”.

I also very much enjoyed living with her at the time despite these foibles. This is not a story about how terrible she was as a flatmate.

2. Unlike the chicken.

3. Although to some degree they probably are this too, because they lack a basic feedback mechanism required to get good at it.

4. Depending on how bitchy I'm feeling sometimes I'd get them to do it instead, but that's often more stressful than doing it myself. This is not a good strategy and I don't endorse it, but it is the one I tend to follow.

5. This is less necessary with library code because the code is the user interface and you can see it being used in the tests.

6. Except in the sense that sometimes you'll find problems and have to fix them.

7. Bad news! I will notice. Yes, including that time you thought you got away with it.