Sunny

← All weeks

What I’ve been reading (W26, 2023)

Stay a beginner · On bad (programming) advice

Stay a beginner

  • To influence a field, it’s not about being the “expert” or “thought leader”; it’s about visibly operating from a base of failure, ignorance, and humility.
  • People won’t learn much from “experts” that have it all figured out; they learn from people they identify with. People see themselves as blind, ignorant beginners like them, charting their path, making one idiotic mistake after another. Similar to great storytelling, they follow the character they identify with, a champion from their ranks taking on an activity they wish they could.
  • The beginner is constantly finding the newbie problems the expert takes for granted. The expert doesn’t need a map and may struggle even drawing one, so familiar they are with the terrain. Beginner knows exactly how to draw the map they wish they had. They know how to give directions and why they’re the best due to painful trial and error.
  • The beginner finds the real first principles through painful failure and humiliation. By finding real first principles, beginners stumble upon true innovation. They experience the genuine problems felt by the user base. This lets them build meaningful and disruptive products connecting real user problems and first principles.
  • If you can stay a beginner, you will have mastered true “expertise” - the ability to use or discard your weak experience as needed. To be a visible presence in a field. Or to discard the field itself as a useless construct meant to protect the status of experts. To master learning itself, not the field.

On bad (programming) advice

  • ‘Programming’ covers an enormous range of activities with different problem domains, team sizes, management structures, project lifespans, deployment sizes, deployment frequencies, hardware, performance requirements, consequences for failure etc. We should expect it to be incredibly rare that any given practice is appropriate across all those contexts, let alone that we could discover general best practices from the outcome of a few projects.
  • Programming practices are mostly tacit knowledge. Tacit knowledge isn’t easy to share. An expert will relate some simple-sounding rule of thumb, but then grilling them on specific cases will quickly uncover a huge collection of exceptions and caveats that vary depending on the specific details of the situation. These are generated from many, many past experiences and don’t generalize well outside of the context of that body of experience.
  • Trying to apply the rule of thumb without knowing all the details tends to result in failure. Phrases like “don’t repeat yourself”, “you aren’t going to need it”, “separation of concerns”, and “test-driven development”, etc. were originally produced by somebody of valid experience but then wildly over-generalized and over-applied without any original nuance.
  • Advice like “write short functions” is a technique that may help achieve that goal in many situations, but it’s not a goal in itself. And yet some characteristic of human thinking makes it very easy for these sorts of pseudo-goals to take over from the actual goals. So you may hear people saying that some software is bad because it has very long functions, even if the evidence suggests that it is also easy to maintain.
  • When means and ends are confused, it’s also often accompanied by the word “should” and other similar phrases that carry some moral or hygienic weight (e.g. “the right way to do it”, “clean code” vs “code smells”). This strips away any consideration of the context or goals of a specific project.
  • Technology companies often make incredible profits using simple technology to solve business problems. This means that programmers at those companies can make many bad decisions, adopt poor practices, and still be successful. If money falls out of the sky, whatever you do, there isn’t much of a gradient to help discover better practices.
  • [You] find much more value in specific experience reports or in watching people actually working, as opposed to writing about general principles.
  • The goal is always to write a program that solves some problem, and that can be maintained over its useful lifetime.

Mental liquidity

  • Mental liquidity is the ability to quickly abandon previous beliefs when the world changes or you come across new information.
  • So much of what people call “conviction” is actually a willful disregard for facts that might change their minds. It’s dangerous because conviction feels like a good attribute, while it’s opposite – being wishy-washy – makes you feel and sound like an idiot.
  • We laugh at who we were at age 20 but assume that by age 60, we’ll roughly be the same person we are today. Part of the reason it occurs is that it’s too painful to accept that the beliefs I hold today might be wrong, temporary, or subjective.
  • For a lot of things in life – particularly politics, investing, and relationships – people don’t necessarily want the truth; they want certainty. Changing your mind is hard because it’s an admission that the certainty you once thought you held was an illusion. The path of least resistance is to cling to beliefs for dear life.
  • Be careful what beliefs you let become part of your identity. Religion and politics are contentious because, almost by definition, your beliefs are part of your identity – you’re not just dealing with ideas and philosophies but tribes and belonging.
  • Most fields have lots of rules, theories, ideas, and hunches. But laws – things that are unimpeachable and cannot ever change – are extremely rare. Some fields only have a handful. A big problem arises when you force rules and theories into laws. The few laws tend to be the most important things in any field. But everything else, like Einstein, said, is just a theory of maybes.
  • Ask, “What have you changed your mind about in the last decade?” to understand how people process new information.

Others

Substack edition

View the original post

Stay a beginner

  • To influence a field, it’s not about being the “expert” or “thought leader”; it’s about visibly operating from a base of failure, ignorance, and humility.

  • People won’t learn much from “experts” that have it all figured out; they learn from people they identify with. People see themselves as blind, ignorant beginners like them, charting their path, making one idiotic mistake after another. Similar to great storytelling, they follow the character they identify with, a champion from their ranks taking on an activity they wish they could.  

  • The beginner is constantly finding the newbie problems the expert takes for granted. The expert doesn’t need a map and may struggle even drawing one, so familiar they are with the terrain. Beginner knows exactly how to draw the map they wish they had. They know how to give directions and why they’re the best due to painful trial and error.

  • The beginner finds the real first principles through painful failure and humiliation. By finding real first principles, beginners stumble upon true innovation. They experience the genuine problems felt by the user base. This lets them build meaningful and disruptive products connecting real user problems and first principles. 

  • If you can stay a beginner, you will have mastered true “expertise” - the ability to use or discard your weak experience as needed. To be a visible presence in a field. Or to discard the field itself as a useless construct meant to protect the status of experts. To master learning itself, not the field.

On bad (programming) advice

  • ‘Programming’ covers an enormous range of activities with different problem domains, team sizes, management structures, project lifespans, deployment sizes, deployment frequencies, hardware, performance requirements, consequences for failure etc. We should expect it to be incredibly rare that any given practice is appropriate across all those contexts, let alone that we could discover general best practices from the outcome of a few projects.

  • Programming practices are mostly tacit knowledge. Tacit knowledge isn’t easy to share. An expert will relate some simple-sounding rule of thumb, but then grilling them on specific cases will quickly uncover a huge collection of exceptions and caveats that vary depending on the specific details of the situation. These are generated from many, many past experiences and don’t generalize well outside of the context of that body of experience.

  • Trying to apply the rule of thumb without knowing all the details tends to result in failure. Phrases like “don’t repeat yourself”, “you aren’t going to need it”, “separation of concerns”, and “test-driven development”, etc. were originally produced by somebody of valid experience but then wildly over-generalized and over-applied without any original nuance.

  • Advice like “write short functions” is a technique that may help achieve that goal in many situations, but it’s not a goal in itself. And yet some characteristic of human thinking makes it very easy for these sorts of pseudo-goals to take over from the actual goals. So you may hear people saying that some software is bad because it has very long functions, even if the evidence suggests that it is also easy to maintain.

  • When means and ends are confused, it’s also often accompanied by the word “should” and other similar phrases that carry some moral or hygienic weight (e.g. “the right way to do it”, “clean code” vs “code smells”). This strips away any consideration of the context or goals of a specific project.

  • Technology companies often make incredible profits using simple technology to solve business problems. This means that programmers at those companies can make many bad decisions, adopt poor practices, and still be successful. If money falls out of the sky, whatever you do, there isn’t much of a gradient to help discover better practices.

  • [You] find much more value in specific experience reports or in watching people actually working, as opposed to writing about general principles.

  • The goal is always to write a program that solves some problem, and that can be maintained over its useful lifetime. 

Mental liquidity

  • Mental liquidity is the ability to quickly abandon previous beliefs when the world changes or you come across new information. 

  • So much of what people call “conviction” is actually a willful disregard for facts that might change their minds. It’s dangerous because conviction feels like a good attribute, while it’s opposite – being wishy-washy – makes you feel and sound like an idiot.

  • We laugh at who we are at age 20, but we assume that by age 60, we’ll roughly be the same person we are today. Part of the reason it occurs is that it’s too painful to accept that the beliefs I hold today might be wrong, temporary, or subjective.

  • For a lot of things in life – particularly politics, investing, and relationships – people don’t necessarily want the truth; they want certainty. Changing your mind is hard because it’s an admission that the certainty you once thought you held was an illusion. The path of least resistance is to cling to beliefs for dear life.

  • Be careful what beliefs you let become part of your identity. Religion and politics are contentious because, almost by definition, your beliefs are part of your identity – you’re not just dealing with ideas and philosophies but tribes and belonging.

  • Most fields have lots of rules, theories, ideas, and hunches. But laws – things that are unimpeachable and cannot ever change – are extremely rare. Some fields only have a handful. A big problem arises when you force rules and theories into laws. The few laws tend to be the most important things in any field. But everything else, like Einstein, said, is just a theory of maybes.

  • Ask, “What have you changed your mind about in the last decade?” to understand how people process new information.

Others

Originally on Substack