Half of my twitter feed was all up in arms over a thread about 10x programmers this week. I am purposely not going to repost the thread or the comments because I don’t think any good will come from it, but essentially it turned into pitchfork style mob criticizing the author for encourage a toxic work environment.

The Myth of the 10x Programmer

In early silicon valley lore, legend speaks of a mythical 10x programmer, that could write code faster than 10 of their piers. They single handedly slayed the dragon and defended the city… or something like that.

Like all myth I think this is actually based in reality.

Around 20 years ago when I first got into the industry there was a bunch of large changes that really changed the way we program in industry. Before Linux and the rise of OpenSource/GitHub knowledge wasn’t really shared between organizations. Even the most trivial things were treated as trade secrets. You couldn’t just learn something - you had to reverse engineer it or be taught about it.

At that time the dotcom bubble was in full force, all needing programmers. More programmers than there were. So startups were hiring people without any programming experience at all. In some cases that worked out really well and other cases it didn’t. The result was that we ended up with a bunch of people in the industry that weren’t setup to succeed - they lacked training and mentorship and never really got great experience.

When the dotcom bubble burst it made a mess of the industry. And this gave way to the startup scene. Startups were coming up with crazy disruptive ideas. Some good, some evil.

And then all of a sudden we were trying to share everything. You could read other peoples code. Share ideas. Collaborate with people you’d never get a change to actually meet in person.

As we got better at sharing code the whole industry seemed to level up. I remember working with C# for the first time on a project and thinking that code that used to take months to write could be accomplished in weeks. And Eventually I found Ruby/Rails which made development (well, lets say prototyping) even quicker.

During this age of enlightenment the myth was born. Back then I think the term that was popular was “Rockstar” developers. I actually think that name makes sense, in a “rock the stadium and then get drunk and trash the hotel room” sorta way.

But the reason why these developers seemed to be so highly productive was a combination of things:

1. They were often taking on loads of technical debt. (And while this is often bad, in some cases this really paid off for them.)
2. Everything was simpler back then, in bad ways. They didn’t worry about security or privacy.
3. It possible to fit a whole codebase in your brain, especially if you wrote most of it. The cognative load was much lighter.
4. New tools made writing code faster.
5. They found ways to cheat that we no long find acceptable. (And probably shouldn’t have then.)
6. They had some good ideas.

What made them look like a 10x programmer or not, depended on how they interacted with the rest of the development team. The people who I think of as great developers lift the level of the whole team and help everyone achieve more. The contribution of someone like this isn’t always obvious and sometimes you only really see it retrospectively.

There were definitely developers at that time who were self centered and only cared about delivering code. And if you were on that kind of team it was probably dysfunctional making the “Rockstar” developer look great and the rest of the team look like they couldn’t keep up. Depending on who you talk to there are tonnes of stories about developers like this. Sometimes they were hugely successful and sometimes they weren’t. It depended on a lot of things and the external factors needed to produce them or make the successful don’t exist anymore.

In general I think we’ve also learned to value different things since then. We don’t like “clever” code anymore - readable code is much preferred. Move fast and break things doesn’t work anymore. People are important. Being inclusive is important.

Writing Software is like Writing Music

Software development is a bit of a creative art. It can be formulaic but it also requires creativity to solve problems.

I think of it like the difference between knowing how to play a guitar vs. being Tom Petty.

You can’t look at someone’s physical appearance and predict their musical ability. You have to hear them play. Listen to their work. And event then that only tells you if they know how to play music - writing music is entirely different.

Writing a song I think is a pretty good analogy for software. You require some amount of base knowledge to do either: how to play an instrument/how to write code. But there’s a difference between strumming some chords and writing a song just like there’s a difference between typing in some code and writing a program.

Rockstars

I think where the talk of 10x developers or Rockstars becomes really problematic is when someone tries to describe a person in terms of their characteristics.

Imagine if the checklist for a Rockstar was:

• known to destroy hotel rooms
• into hardcore drugs
• constantly drunk
• self destructive
• plays a specific brand of guitar

That’s crazy right. That doesn’t even involve music. You end up with all sorts of disaster.

But that’s really where we are when we talk about 10x programmers. We’re looking for something that can’t really exist anymore based on characteristics that have nothing to do with their ability.