Reed Hastings: On rock star developers

Many of the developers I admire most, if they be rock stars at all, are rock stars in the Neil Peart mold: they're unassuming and don't take center stage, but are widely regarded as geniuses and masters of the craft, setting the vision and structure of the project. Good examples include John Carmack (who was perfectly willing to let Romero engage in rockstar theatrics while he hacked out the next Id engine) and Dan Ingalls (Alan Kay gets most of the glory for Smalltalk, but even he would say Ingalls did the lion's share of implementing the thing).

I like to call myself a roadie developer, swooping in to clean up the mess after the rockstar developer came in and trashed the stage...

Everyone seems to be trashing the idea of 10x developers (and I would tend to agree unless you’re literally talking about a 1 in a million mythical dev) but there are definitely 2x or 3x developers I have worked with at companies.

Unfortunately, I can’t imagine a team full of just these developers (in my personal experience) as these folks tend to be bossier than the rest and/or have true disdain for non-programming work which I can empathize with, but is essential in team settings.

Netflix is famous for paying top salaries, but do they actually pay 10x the average salary? Is the rockstar getting a raw deal at Netflix?

Rockstar developers can duct tape a bunch of bullshit together and make a good demo, and after 6 months when they get poached by another company you'll just have to hire actually competent people. They'll have to untangle the your geniuses opus of bullshit and rebuild it into something that works. It will take time and it will be expensive.

I'm not too bothered by people wanting to hire rockstars though because there will always be a market for people like me to come in and clean up the mess.

It's interesting NBC is still covering something like this. I feel like this is generally well-known in the industry, and that any lean company will try to hire a "rockstar" whenever they can, as they don't have the budget to waste on a pretty good engineer (although, I know in recent years myself and many others personally despise that term "rockstar" and think it causes more harm than good).

I don't know what the term "rock star developers" actually mean, but there are definitely people I would call short term and long term developers.

The progress of a short term developers work, looks extremely fast, but its logarithmic. A lot of bad project managers think he does an insane amount, maybe 10x, maybe 100x levels of work. He writes code so fast and cranks out so many features its no wonder he's a rockstar. After a while though, because everything he builds is built for the moment, its just nothing but mazes full of bugs that frighten even the nerdiest dungeon crawling hacker.

The progress of having a long term developer on a team, on the other hand, starts off slow, but it's exponential. He might need weeks or months to build up a good foundation. He spends time writing small things that show little to no benefit, but are built well and from which other things can be built on top of. Everything he builds can be used by anyone who reads it, so they too can use his work. His work multiplies the work of his own as well as the rest of the team. But the bad project manager sees him and thinks he's too slow to complete tickets. He is a .05x developer at this point because the average developer would have been finished with two features by now.

After weeks, maybe months, however, his progress matches those of the average developer. In time progress is so quick, it makes the short term developer's work, which is starting to plateau, look effectively flat. While the short term developer was spending his time building up walls with his hammer, nails, and duct tape, the long term developer was there making a drill, that was used in making a saw, that was used in making a ladder, that was used in making a crane. Now he's making skyscrapers using the tools he built, something the short term will never even be able to do in a single lifetime.

The test can not be conclusive unless all of them worked on same type of projects, programming languages and faced same bugs over the years.

It's quite likely the so called rockstar had a prior experience in the problem at hand. It will be good to see the engineers learn an entirely new language and solve an entirely unrelated problem

The thing to draw out from the study is hire the right fit. There is nothing called rock star engineer. You rockstar engineer is the best fit for your job, tech stack and challenges in your business.

This article makes the dumbest assertions I've heard about quality software development in quite a while.

The "study" isn't really a study. Even so, the metrics that it claims to derive aren't there for the picking.

Then throwing in big names like Bill Gates to say "look Bill Gates said it too" is just name dropping.

Also this looks like something someone would write in 3 minutes between meetings and never actually proof read. The quality of the writing is really quite low.

It's possible to be a rockstar developer without being a diva. Some devs are just genuinely skilled at the tech and also at being good coworkers.

> The researchers expected that the best programmer would outperform his average counterpart by a factor of two or three. But it turned out that the most skilled programmer far outperformed the worst. He was 20 times faster at coding, 25 times faster at debugging, and 10 times faster at program execution than the programmer with the lowest marks.

What? Okay, so how much did he outperform the average one?

I had to laugh at their test. Two hours? Many people can sustain a high pace of work for two hours. But how about as time wears on? There's a lot of assumptions going on here.

stepping back, im wondering what would motivate the CEO of netflix to write on article like this on CNBC?

Anyone have any ideas? It's essentially a blog post.