4 ms·

One nice thing to share here is that a statistical model for success under such conditions is for there to be a collection of independent random variables that

One nice thing to share here is that a statistical model for success under such conditions is for there to be a collection of independent random variables that multiply to form your overall success: people who lack any particular one of them, x_1 and x_2 = 1000 but x_3 = 0, see the product x_1 x_2 x_3 = 0, below people who have x_1 = 10, x_2 = 10, x_3 = 10. Taking a logarithm, one gets a sum of independent identically-distributed variables, which then tend to sum to a normal distribution due to the Central Limit Theorem, meaning that success distributions produced by multiplication tend to be log-normal. This was argued by John D. Cook at [1]. Log-normal distributions are interesting because they have a long tail—usually the top people drawn are way, way better than you might have expected; normal populations stick within several standard deviations of the mean but long-tailed ones see remarkable achievers many tens of standard deviations above the mean. This result has been long known for things like wealth and income and other measures of external success. So there is a lot of truth to this idea that, say, if you are a programmer, your time is worse spent on building that extra expertise with things that you are good at, like programming, and better spent on building new expertise with things that you are not good at, like perhaps digital art or negotiation/persuasion or accounting. [1] https://www.johndcook.com/blog/2009/09/29/achievement-is-log-normal/ https://www.johndcook.com/blog/2009/09/29/achievement-is-log...

wjnc 7y ago
It's quite a stretch to go from an argument on distributions to an argument on the best marginal use of ones time. Those in the right of the log normal curve have had to learn their skill as well. It's all about trade offs. The regular advice I've learned is to improve your primary skill until you are a professional senior, then expand your skills. And even that then goes against the popular Federer story who started in other sports and whose tennis teaching mother wouldn't teach him!
nostrademons 7y ago
This changes with the size of the firm you're part of. It's true for individual pursuits like basketball, drawing comics, or entrepreneurship. But a big corporation gets around a lack of any one critical variable by hiring someone for that variable. So if you're an excellent backend engineer, your time is wasted learning digital art, because somebody else will be making the artwork. Effectively the corporation becomes the organism and you become one specialized organelle in them. You're better off increasing your expertise within that specialty than learning other ones - say by firsthand experience with all the ways distributed systems can fail. The returns to being a top individual producer beat the top returns to being a corporate drones, but the average returns are significantly worse (because the big corporation benefits from being able to swap out individual motivated actors to plug deficiencies rather than having to develop skills itself), so depending on your risk tolerance, this may or may not be a good strategy.
- adrianmonk 7y ago
  > So if you're an excellent backend engineer, your time is wasted learning digital art, because somebody else will be making the artwork. Tangentially, as a backend engineer, I've had to tell the people who made the digital art that the additional icon they supplied me wasn't quite the same color as the existing ones it was supposed to match and also looked misaligned because it didn't follow the same geometrical conventions. So I wish what you're saying were true, but it isn't necessarily. This being a tangent, I'm not necessarily looking for a moral to the story here, but maybe it's something about having a certain minimum competence in an area so you can be sure you're working effectively with others. (EDIT: What is a backend engineer doing with an icon at all? Sticking the URL where it's hosted into a config file and modifying the logic so that it shows up (or doesn't) as appropriate, then testing that by looking at the UI and seeing art that looks wrong.)

wjnc 7y ago

It's quite a stretch to go from an argument on distributions to an argument on the best marginal use of ones time. Those in the right of the log normal curve have had to learn their skill as well. It's all about trade offs. The regular advice I've learned is to improve your primary skill until you are a professional senior, then expand your skills. And even that then goes against the popular Federer story who started in other sports and whose tennis teaching mother wouldn't teach him!

nostrademons 7y ago

This changes with the size of the firm you're part of. It's true for individual pursuits like basketball, drawing comics, or entrepreneurship. But a big corporation gets around a lack of any one critical variable by hiring someone for that variable. So if you're an excellent backend engineer, your time is wasted learning digital art, because somebody else will be making the artwork. Effectively the corporation becomes the organism and you become one specialized organelle in them. You're better off increasing your expertise within that specialty than learning other ones - say by firsthand experience with all the ways distributed systems can fail. The returns to being a top individual producer beat the top returns to being a corporate drones, but the average returns are significantly worse (because the big corporation benefits from being able to swap out individual motivated actors to plug deficiencies rather than having to develop skills itself), so depending on your risk tolerance, this may or may not be a good strategy.

adrianmonk 7y ago
> So if you're an excellent backend engineer, your time is wasted learning digital art, because somebody else will be making the artwork. Tangentially, as a backend engineer, I've had to tell the people who made the digital art that the additional icon they supplied me wasn't quite the same color as the existing ones it was supposed to match and also looked misaligned because it didn't follow the same geometrical conventions. So I wish what you're saying were true, but it isn't necessarily. This being a tangent, I'm not necessarily looking for a moral to the story here, but maybe it's something about having a certain minimum competence in an area so you can be sure you're working effectively with others. (EDIT: What is a backend engineer doing with an icon at all? Sticking the URL where it's hosted into a config file and modifying the logic so that it shows up (or doesn't) as appropriate, then testing that by looking at the UI and seeing art that looks wrong.)