Getting to Ten Times Better
Among the many points made by Nicholas Nassim Taleb’s The Black Swan is that certain things come from power law distributions rather than from normal, bell-curve distributions. Nevertheless, academics (and normal people too) use the elegant, familiar, and well-trodden statistics of the bell curve. One of the points of The Black Swan is that this blows up in their faces.
Among the things Taleb repeated but I was too dense to really absorb without some additional life experience is that almost all the stuff that really matters in life is drawn from the power-law distribution. I would like to review some of these areas in this post.
I am the height of the average American male. In contrast, our esteemed Executive Editor Will Truman is very tall. Even so, he is not ten times taller than me. He is only perhaps about 15% taller than me. Despite this small percentage difference, it is enough to put him on the far end of the bell curve. A humble 15% difference is enough to make Will an outlier.
Let’s contrast that with something more complicated and, frankly, more interesting: basketball performance.
I wouldn’t be terribly surprised if Sam Wilkinson were five times better at basketball than me on the rare occasions when his ankle is functional. Even if he’s not, there are thousands of professional basketball players in leagues around the world. Every one of them is probably ten times better than me, and a healthy chunk of college players would also make the list. Some of them are even both shorter than me and ten times better players than me. Most of them are likely much taller than me though.
Famously, height tends to be helpful in basketball. It’s one of the many inputs that feed into ultimate basketball performance. Even so, it cannot explain 10x differences. Will might not be a better player than me at all despite his height advantage. To my knowledge, he hasn’t invested himself in the dozen or so associated skills and properties that would make him a dominant player. It’s these investments, along with a height advantage, that can make an elite player.
There are many articles online about the 10x programmer. There are just as many or more about the myth of the 10x programmer. The idea is that there are some “hackers” who are just so good that they are ten times more productive than the average programmer one can hire. If you believe in 10x programmers, even if the average programmer salary is, say $150,000 per year, it makes sense to do an extensive search for a 10x programmer and pay them $300,000. You will still get a 5x benefit in productivity.
The war over the existence of 10x programmers is an important one. It may be responsible for some of the extreme salaries that software developers can achieve at certain companies.
But do they actually exist? Or is it just a way to justify paying someone $300k?
I suspect that parallel to the height-basketball world, that there are comparatively modest differences in the ability of different programmers to perform core, elemental tasks. For example, among programmers who know what bubble sort is and how it works and is able to implement it at all, the slowest programmer will not be ten times slower than the fastest programmer. Rather, we will see completion times that seem drawn from a bell curve without dramatic differences. In fact, this is what programming classes often rely upon. Students are given exams where they need to perform some familiar task. Some students are fast and others are slow, but 10x differences in the performance on small tasks are unlikely.
This changes when one goes to project work. As problems become more complicated, they require a broader array of skills. If a student is strong at all of them, they work together to allow her to complete her work straightforwardly. Meanwhile, a student who is weak at all of them could foreseeably take five times longer.
Real life is closer to project work. There are any number of architectural choices that one could make at the outset of a project that have downstream effects that either make subsequent development a breeze or a death march.
What is a programmer’s job really though? Are they given well-defined projects to code? Some are, but many have some degree of influence over what they do and in what order. Making intelligent choices about what gets completed in addition to what way can easily lead to 10x differences in realized productivity. In fact, a mediocre developer who chooses to work on an important business problem that impacts customers may be 10x or 20x as useful than a super-smart developer who works on doing a superb job on something that doesn’t actually matter.
Interestingly, people who are skeptical of the existence of 10X programmers are often extremely convinced of the existence of a 10X+ dynamic range in the productivity of firms.
— Patrick McKenzie (@patio11) February 17, 2020
How does one become a billionaire? I haven’t quite succeeded at this myself, so suspect my advice.
Nevertheless, I doubt any answer that starts with “work hard.” Working hard might indeed be necessary, but I suspect that it’s like advising a young person wanting to join the NBA to “be tall.” As Ezra Klein tweeted yesterday, lots of people work hard:
"I worked very hard for it" is the worst possible justification for being a billionaire. You know who works very hard? Single mothers working two minimum wage jobs with just-in-time scheduling and a bus commute.
— Ezra Klein (@ezraklein) February 20, 2020
Working hard is not sufficient. But this isn’t the only problem with the advice. Another problem is that how hard you work is definitely drawn from a bell curve, distribution, not from a power-law distribution. It would take a Herculean effort for a person to work twice as hard as the average person in that profession and situation. A billionaire, however, is worth 1000x a millionaire. You’re not getting there by showing up to work an hour early. You are only getting there by producing something 1000x more valuable than the millionaire.
What makes something valuable? Value is in the eye of the buyer. The buyer doesn’t care how much effort you put into it. They want something that solves their problem.
If you want to become a billionaire, you must know that no one will care if you pour your heart and soul and suffering into your work. That’s what artists do, and the median artist makes nothing. The buyer wants something that can solve their problem.
I suspect that to become a billionaire, you need to explicitly decouple any mental connection you have between how hard you are working and how much you are making. Rather, you need to focus maniacally on producing what your customer is willing to pay for. You might need to go further and consider those things multiple customers may be willing to pay for; to become a billionaire likely requires you to provide value to many, many people, perhaps most of whom you have never have direct contact with. In the course of doing this, you might find yourself working hard at times, but this is a by-product of trying to produce value. To put the hard work first on your route to billionaire-hood is backwards. If you try it, you will likely find yourself stuck working hard on something that you find satisfying on some craftsmanship level but doesn’t matter to your customer.
Nothing in this post tells you how to actually get these sorts of improvements in any of these domains. I suspect that the moves required are sufficiently subtle that they are hard to articulate even by those who have done it. Nevertheless, Taleb’s point holds: many of the things most important to us belong to these power-law distributions. Finding out how to get on the edge of one of these curves is worth the effort.