Sunny

← All weeks

What I’ve been reading (W43, 2022)

Variability, Not Repetition, is the Key to Mastery · Ideas are the Enemy of Observation

Variability, Not Repetition, is the Key to Mastery

  • Variability plays a role in abstracting the deep structure of seemingly different situations.
  • Mix up your problems. If you have to study problems, mix them up so there aren’t clues telling you what strategy you need to apply. Mix together Unit One and Unit Two—and don’t indicate which problems come from which sections.
  • Choose opportunities with greater variability. Management consultants who work for a wide range of clients are likelier to learn broad principles than those who work only with a single firm or industry. Psychologist Gary Klein reports that inner-city firefighters progress much faster than rural firefighters, owing to the more varied firefighting conditions. Choosing jobs with greater variability early in your career may lead to better skill development than narrow specializations.
  • Work with multiple teachers, peers and styles. Learning a language, for instance, benefits from exposure to different speakers, accents and speaking styles. Learning from several teachers is more likely to promote diverse perspectives. Having a variety of peers exposes you to different strategies and is less likely to get you in a rut.
  • Add constraints that force you away from dominant strategies. Performing skills under different constraints force you to explore wider areas of the problem space This added variability helps you discover new strategies and kicks you out of ruts that can develop when an easy, but sub-optimal, method is relied on.

Ideas are the Enemy of Observation

  • An idea is a set of smug certainties that allow you to stop looking, listening and learning. Observation is the antidote to those certainties. Ideas are rigid, and observations adapt. Ideas make you seem smart, and observations make you smarter. But for a writer, the most important distinction is this: Ideas are generic, and observations are specific.
  • Every day, try to write down ten observations. Then re-read them and ensure none of them carries the tainted whiff of your ideas. Write down what you see and hear on the street, not what you expected to see and hear, and not what you presume is going on. This is hard. At first, all you will see are things that confirm your preconceived notions.
  • When it comes to changing the world, nothing is more powerful than a truthful observation. If you want to take on the meat industry, you don’t write a healthy-eating manifesto; you write “The Jungle”. If you want to say something meaningful about race, don’t pile up a bunch of high-minded, heavy-handed parables, like in Crash; pile up a ton of true-to-life observations, like in “Homicide” or “The Wire”.
  • Ideas, I now see, are the true recipe for passivity, and observations are the true spur to action. But you can’t observe anything if you’re using your ideas as an excuse not to pay attention. The worst bias a writer can have is confirmation bias.

A high production rate solves many ills.

  • Any given technology development is how many iterations you have and what’s your time and progress between iterations.
  • In software, the production rate is the iteration rate. If you have lots of iterations, it’s OK if one of them is a potentially high-value experiment that doesn’t pan out. If you have one iteration per year, you are less likely to want to take that risk, and your reluctance will be justified.
  • The higher iteration rate allows you to take more risks, which in turn allows you to push the boundaries more and thus gather more relevant feedback in each iteration, at the same time that the reduced time frame reduces what you can do.
  • If you iterate and use the feedback the iteration gives you; you will improve each time because you will have learned something.

The Perfect Commit

  • Our job as software engineers is not to write new software: it’s to apply changes to existing software.
  • The commit is our principle unit of work. It deserves to be treated thoughtfully and with care.
  • Each commit should change a single thing. Atomic commits are also much easier to revert if something goes wrong cleanly—or to cherry-pick into other branches.
  • The ultimate goal of tests is to increase your productivity. If your testing practices slow down, you should consider ways to improve them.
  • If your project defines APIs meant to be used outside of your project, they must be documented.
  • Documentation is only valuable if people trust it. People will only trust it if they know it is updated.
  • Every perfect commit should include a link to an issue thread accompanying that change.