The concept of the 10x engineer is an incredibly popular one to debate. People love to write about it, like how to become one, their demise, what they really are, and all of the dimensions of one. This is kind of like one of those posts, except I’m going a step further and saying that not only is it obviously true that 10x engineers exist, it’s also true that 1000x engineers exist. That is, engineers that create 1000x more value for a business than the “average” engineer. People like Jeff Dean, John Carmack, and Steve Wozniak are way more than 1000x the average engineer, so it’s a mystery why we’re still debating their existence.
So what does a 1000x engineer look like? There are three roughly independent dimensions in which engineers can be 10x the average: problem solving ability, work ethic, and business context. If you find people that are multiple standard deviations above average in each of these, they will generate incredibly high returns for your business. Other stuff like communication is often brought up, but there is less potential for 10x there and it’s strongly correlated with business context. Let’s look at the three dimensions more closely.
Problem solving ability is most closely tied to raw intelligence and usually the most controversial one. But if you’ve programmed with people for a while, it’s easy to observe vast differences in people’s ability here. My favorite test of this is complex debugging problems, which involve your brain exploring a massive search space and pruning out branches until you get to the answer. Great problem solvers can debug well over 10x more quickly than the average engineer (who is typically terrible at debugging), and debugging is the hardest part of programming.
Next is work ethic, which is a combination of personal and environmental factors. It’s pretty unlikely for someone to work 10x as many hours as someone else, but if you work 3x as much (so maybe 60 “real” hours vs 20 per week) and let the benefits of that work ethic accumulate over many years, you end up with a much bigger factor. People with 10,000 valuable hours of programming are incredibly good at what they do — it looks like amazing intuition about how to approach problems. Engineers often get a bad rap for working too much, but it really accelerates your valuable experience.
Finally, the least controversial one is business context. Most of time wasted in engineering is time spent working on the wrong thing. A product that nobody wants, tools with no purpose, overly complex technical problems for their own sake. Smart engineers fall into this trap all the time, especially at big companies due to “failure” of their manager or leadership. At a startup, you get relatively quick feedback on whether what you’re doing is useful at all, so using that feedback and making good engineering decisions yields massive returns. It’s not crazy at all to 10x the probability that your startup succeeds from 0.01% to 0.1% doing this — but not all startups have this property. It’s tough to be 10x on business context if your idea’s success depends on government regulations, for example, so the opportunity matters a lot here.
Between these three 10x factors we have 1000x. There are a lot of implications to believing that 1000x engineers exist. If you have a 1000x engineer work with a 1x engineer (or even a 10x engineer), you’re wasting their time. If you don’t design your organization to support these people, you’re probably not as successful as you could be. If you’ve worked with a 1000x engineer before, you’ll be desperate to work with someone of that caliber again. Maybe this is all stretching reality a little bit, but I think it’s closer to the truth than most people want to believe it is.