Hacker News new | past | comments | ask | show | jobs | submit | page 2 login
Things I’ve learned in my 20 years as a software engineer (simplethread.com)
1605 points by D_Guidi 3 months ago | hide | past | favorite | 722 comments

The 10x programmer is a silly myth. The idea that someone can produce in 1 day what another competent, hard working, similarly experienced programmer can produce in 2 weeks is silly.

You know, 10x is an optimistic number here. Some programmers will do in 1 day what you wont achieve in a life time. And not understanding that makes you a bad programmer by my book simply because this is the foundation of the job.

So let me give one advice: you are not producing code, you are engineering code. Some will find solutions that the others will not find. Thats where the 10x programmer thing come from. The guy who wrote bittorrent said something like the following: "I've written bittorrent in a way you would never have thought about".

You think John Carmack was the only one trying to make 3d games when Doom came out? Thousands of programmers where trying to. Does that make him a 10x programmer in your opinion? Or more like a 1000x?

Agreed, I don't know why the industry (or people?) in general have a problem with accepting this abilities gap.

It's as if the notion of people being arbitrarily born more or less with a knack for something is evil.

Let me tell you the story of when I realized I was (at best), a "not-mediocre" dev:

  1. I've been programming for fun even as a kid. That's what got me into it: If you count kid-programming as "programming", I've been programming for ~14-15 years.

  2. There was a girl who used to work with me at $DAYJOB, who is fairly well-known on The Orange Website, and has written some "thought leader" papers in certain niches. A real einstein.
I asked her in DM at work about her background and how long she's been coding -- both in general and the language she's well-known for. The answer was:

  "A couple of years -- and a few of those were writing another language. This is my first real job writing $LANGUAGE_WELL_KNOWN_FOR."
Well, fuck me.

This is called "innate ability". Humans aren't created equal, nobody said they had to be, and it's not a crime to acknowledge it. Folks ought to get over it IMO, it's not a big deal.

I will never understand CS and programming-language theory concepts at that caliber or that intuitively, but who cares?

I'm also short as fuck so I doubt I'll be playing in the NBA anytime soon either. Life goes on.

There was some tweet recently that went "Saying 10x engineers don't exist is as true as saying 10x basketball players don't exist or that all writers are basically the same." It's self-evidently not true.

No. It's like saying that a good mechanic can fix a car 10 times faster than another. Even if a good mechanic manages to find the problem in less time than a non experienced one, he still has to do the repair.

That is similar to programming: even if you have everything clear in your head, you know how to do every single thing that is written in the requirements of the program, you still have to type it out.

A lot of time I thought, well this is fairly easy, and then spent I month just typing out code.

The only situation where really a more experienced programmer can be more or less 10 times faster than a non experienced one is in debugging/troubleshooting problems, because with more experience he tends to know by experience where it's most likely that the problem is and thus test in that direction.

You are misunderstanding the definition. A 10x developer does not write code ten times faster than average. They have ten times higher impact on the results of an organization than average. And that's not far-fetched at all. This does not magically happen. There are individuals behind it, often a surprisingly small number doing the critical work that visibly brings the world forward.

It's not about writing a regular boring CRUD app that no one cares about as fast as 10 regular developers. It's about choosing the right problems to solve on their own initiative, building the foundation of a technology organization that can sustainably & quickly grow in scope and users, being the first to solve a problem no one else has understood such that the organization crushes their competition and so on.

Leading the development of web software that scales to 100 million users in 2003. Planning novel computer vision algorithms to a vehicle fleet of millions of vehicles, running on custom hardware they've also made contributions towards. Realizing that consumer VR headsets are now _just barely possible_ if you do everything right, and bringing it to market. Etc.

This is the difference between a small organization being capable of solving a difficult business problem and achieving profitability, and not.

F1 mechanics changes tires in 2 seconds. For my local workshop it takes an hour. That's like 1000x.

Bad analogies aside, as others already pointed out, the biggest gain is usually on more creative and architectural tasks with amplified impact, where the requirements given are not as clear as code. Not how fast you change a bolt or type 100LOC. Instead, for example choosing the right tools instead of reinventing your own, or writing a composable interface that enables integration with other tools instead of of an unmaintainable spaghetti monstrosity.

I'm not sure that fixing something that is broken is a good allegory for what programmers do, because a lot of us (most of us?) create new things.

The impact that creating something new can have can range from changing the course of history, to literally nothing (like most everything I've written).

Programmers and mechanics have a vastly different impact potential than one another.

No: One developer probably cannot _type_ x10 as fast as another. But they might be able to write code that can cure cancer -- while you/I can only regurgitate the same CRUD app over and over.

IMO having a fancier keyboard that does more clickety-clack will not make you a better programmer. If anything, your co-workers cannot work as good and will have to put more energy to focus.

Experience and the skill to remember or recall well, help a lot with engineering.

A fancy keyboard on its own is probably not a good indicator.

But I would switch this around and definitely say with my few years of experience, those that were “the best” were also those that were interested the most in their tools of the trade.

And whether that is a mechanical keyboard or a better screen or better performance or better IDE setups. Doesn’t really matter. Sadly many companies don’t seem to encourage this enough imho

sorry - you'll have to pry my model m from my cold, dead fingers

> Agreed, I don't know why the industry (or people?) in general have a problem with accepting this abilities gap.

I don’t think it’s that people have a problem admitting that there are significant outliers. It’s that there’s a backlash to the fetishization of “hiring 10x programmers” and most people claiming to be one really aren’t.

I know I’m good at it, but I am not and will probably never get to 10x on just writing code. I’ve worked with people who have a better claim to being a 10x. And having seen that, I have also never encountered any company that claims to hire 10x engineers that actually has a process to consistently hire even 1-2x engineers.

If there are 10x engineers, it's a right place, right time situation. Very good engineers with the right experience, luck, a good idea, and the relentless desire to execute. Miss any of those things, and you can still be a a multiplier, just probably not 10x.

Even so at my companies the difference between between good engineers and bad is massive (a lot of the difference is talent and motivation). Imagine you had a company full of shitty engineers, a motivated good engineer in the right could deliver 10x the value of some shit engineer

There's honestly quite a few developers that produce zero or negative value (e.g. by implementing the wrong things, introducing lots of bugs or significant complexity etc). 10x:ing them is not difficult.

Totally agree with you. I’m in the same category and have had to accept my fate to some degree. I keep improving relative to myself, but that’s all I can reasonably do.

What gets me is when I meet many of these “innate talent” coders I get to hear phrases like:

“I just want to work with the best possible team…”

And then from their behaviour, you see how they get irritated with the regular folks and try and convince everyone that if we all just “be more like them” it’ll all work better.

Many of them are blind to how much their talents make what is easy for them, very hard for the rest of us.

I call this the “talented arrogance” syndrome.

> Many of them are blind to how much their talents make what is easy for them, very hard for the rest of us.

See also: Asking very old people how they managed to win so long.

Regarding this issue of "innate" talent, I want to ask, how do you know?

there are just so many variables that go into human cognition. Belief. Drive. Focus. Grit. Exposure to ideas. Social support. Physical resources. Exercise. Health.

Each of these contribute, in some way, to every moment we spend learning. They are always changing. Each moment we learn affects the outcome of the next. It's Conway's game of life in n-dimensions.

My second question, and this is perhaps more important; what is the effect of a belief in innate talent? What don't you do or try because you believed it wouldn't work? Because you believed you couldn't? What do you beliefs cost you?

I realize my questions are pointed, and for that I apologize. They are, however, genuine.

They've done some pretty extensive studies on ability / achievement. The big driver being parents wanting to ensure their kids turn out great, or to make their kids smart.

Turns out environment can only really have _detrimental_ effects on outcomes for kids. You can't increase what they don't have, but you definitely can fuck them up. I thought it was pretty interesting.

So basically, the advice most parents get now is, don't drop them on their head :) Of course, that also means provide them with the opportunities to learn / use their ability.

Can you link to these? I'm interested in what they found out, if a little skeptical about how thinking can be precisely measured.

I went back and dug to find the exact source I was thinking of (I heard it in a freakonomics podcast long ago, took awhile to find).

https://freakonomics.com/podcast/reasons-to-be-cheerful/ https://www.inquirer.com/philly/health/Nature_vs_Nurture__Do...

The big thing is this is purely about IQ.

Think of it like folks being born at a certain IQ, and you can really only maintain it (exercise, love, healthy eating, sleep, etc.) You can't bolster it beyond what is innate. However, you (and your environment, typically our parents) can definitely cause decreases in your IQ (poor sleep, anxiety, poor nutrition, toxins like lead, etc.)

I found it pretty interesting, but also kind of obvious in hindsight. I would love to be able to increase my IQ, as would many others, but we all gotta just work the best to maximize what we got. Maybe tech / gene manipulation will make it possible in the future though.

So, there are a ton of resources out there on how unreliable IQ is, and whether it's a valid indicator of general cognitive ability. Instead of linking them, I'll just link to the place that has the most good links: https://en.wikipedia.org/wiki/Intelligence_quotient#Reliabil...

My understanding is that save for general statistical measures of a large aggregate, IQ is a poor measure for an individual (as it has been shown to change quite a bit for individuals over the course of a lifetime).

It’s probably not important whether the talent is encoded in someone’s DNA or whether it was formed during their childhood and early education they got. But by the time people are teenagers, there can be large differences in learning ability that are impossible to level anymore. You can push through those differences through “belief, drive, focus, grit,” etc, but the baselines are different.

there are differences between any two people anywhere ever, though. Is there some reason that being over 20 means you're done developing?

You can learn a new foreign language when you’re 20. But it’s way easier when you’re 10 or 5. Same goes for learning to play chess, to code, and more broadly, for learning to learn.

I'm 35. I started learning to code 5 years ago. I have a job in the industry, and I advance my understanding of the space daily - I'm sorry, but that just doesn't gel with the actual, firsthand experience I have with this stuff.

You're never past expanding your understanding of the world.

Isn't this because everything is new to you when you have close to no life experience? Aside from our default programming mode of eat, sleep, walk the bucket we call brain is practically empty of experiences and theories when we are 5.

> Agreed, I don't know why the industry (or people?) in general have a problem with accepting this abilities gap.

Bias and poor processes. When people become reliant upon a technique or process to attain viability that means of execution becomes more important than the thing you are producing. As a result people will defend slow unproductive means of execution to the death. This is also a socially reinforced phenomenon.

It is completely realistic to achieve 10x (or even much greater) productivity and it isn’t challenging, but it requires abandonment of closely held ideas that are socially reinforced.

It's harder to achieve 10x in a vacuum, but if you make 10 engineers twice as productive, there ya go

> It is completely realistic to achieve 10x (or even much greater) productivity and it isn’t challenging, but it requires abandonment of closely held ideas that are socially reinforced.

OK, you piqued my curiosity. Can you elaborate?

This is completely off the top of my head.

Objective quick wins:

1) Less code. The less there is the less there is to debug and maintain. This applies to all code including dependencies, frameworks, abstractions, and decoration in your own code. If abandoning OOP means substantially less code then do it without question.

2) Iterate faster. If you have a build that takes 10 minutes and the next guy has a build that takes less than a minute they can fail 10x more frequently and achieve the same level of productivity.

3) Know your foundations with confidence. Things you rely upon become more of a barrier than a help given a long enough time line (shorter time line that could otherwise imagine).

4) The point of software is automation. If you find you are doing manual curation in the code you are likely wasting your productivity. An investment in better automation returns interest on the initial investment.

More opinionated personal preferences:

1) Unless you are writing in a language where you manage access to memory, such as C++ or Rust, abandon OOP. In garbage collected languages like Java, JavaScript, C# it only provides composition. It is a tremendous increase in code that does more harm in troubleshooting.

2) Prefer a statically typed language. As a JavaScript developer I have fallen in love with TypeScript. There are performance benefits to ensuring all primitives are strongly typed, but the developer performance benefits come from typing complex types. This is a quick win for defining execution of microservices and functional programming.

3) Prefer functional programming techniques and code portability. Code is less defective when it is highly predictable and explicit.

4) Save composition for last. Don't worry about how to put things together. Complete each micro-task in isolation, which forces a separation of concerns. Once everything is written putting things together becomes natural and typically works itself out.

5) Avoid pronouns in the code. The pronoun of JavaScript is the keyword this. Don't use it. You don't need it. I have been writing in this language full time for 13 years both at work and for personal projects. In that time I have really needed this only twice. The problem with pronouns is that they aren't clear (in either value or intention) by just reading the code, which takes substantially more time to interpret as a human.

6) Think in writing. You can save some refactoring effort by having a solid execution plan. The best way to plan is write things down as if to explain it to somebody who has no idea. If you cannot even explain it to yourself you won't be able to do in code.

Great stuff!

> 2) Iterate faster. If you have a build that takes 10 minutes and the next guy has a build that takes less than a minute they can fail 10x more frequently and achieve the same level of productivity.

I think this is very under rated. There are so many things that can reduce developer latency (time between making a change to the code and seeing if it works, or deploying to production) is one of the most under rated aspects of software engineering productivity.

Most other factors pale in comparison.

Yeah, 10x is really fast. Lots of companies couldn’t bear to go that fast … the rate of change scares them.

The reason people struggle with the idea of 10x programmers is that people are very bad at evaluating the impact of others. Obviously we can all agree on a select few but for those who are just "regular" workers there is often not enough context.

I think it's also because we have a lot of ego. Let's admit it. There are some people who have a rare insight or mobilize people and code in a way that provides 10x value to the company and people.

Think about all the kickass Python libraries people have wrote. Those provide 10x value across the board to entire industries.

> I don't know why the industry (or people?) in general have a problem with accepting this abilities gap.

Because it's not in their interest to accept it openly.

What would be the point of a employer to tell their employees that they are mediocre at best and no matter how hard they try they will never shine? That is the exact opposite of the carrot dangling ever slightly out of reach.

Why would someone blessed by luck tell their employees he didn't get there by working himself to death?

McDonalds is a money printing machine despite not using world class chefs. Telling everybody they can code depending on effort and hiring them for relatively cheap is a solid business plan.

I figure there are a lot of aspects to the job and I'm exceptional at some of the ones most people suck at, and I dont know much about some common things and can find that embarrassing at times. The rest I'm average. So I try to find work where I can bring those less common skills, and I try to fill my gaps too sometimes. Everyone is different, find a place that needs what you have to offer.

I accepted this long ago. I don't have the right brain to be an über-dev, I'm not technically gifted, I just have a high enough IQ to brute-force my way through technical challenges.

I'm great at writing, simplifying technical concepts, and strategizing on larger projects, so I try to work to those strengths, which usually means making my PM's life a hell of a lot easier.

> It's as if the notion of people being arbitrarily born more or less with a knack for something is evil.

The stretch from person A is a better programmer / SW engineer than B to 'thus, person A is innately better' is unnecessary.

In my experience, the programmer that spends more time programming is almost always better than the one who does not. You can hate side projects, open-source work, whatever as much as you want, but IME, the programmers that engage in that (and thus have multiple times more actual experience than one who works 9-5) are universally the 10x people.

In other words, it's not innate intelligence... it's simple effort and working towards improvement.

Now... there are some people that are truly just stellar, and I can attribute perhaps some of this to innate ability, but the difference between mediocre programmer and 10x is typically one of hard work. Between 10x and the extremely rare 100x... well maybe we can talk brain differences. Those programmers are so rare though, it's futile to expect you'd be able to hire them, since there's not enough supply. 10x is good enough.

> You can hate side projects, open-source work, whatever as much as you want, but IME, the programmers that engage in that (and thus have multiple times more actual experience than one who works 9-5) are universally the 10x people.

The usual correlation-vs-causation disclaimer goes here. How many open source developers spend time on it because they produce good code?

I spent quite a bit of time on side projects and helping (as I thought) others with projects that I like. After many years of doing this I have come to the conclusion that I simply cannot produce really good code. So rather than being a drain on already overworked maintainers I decided not to bother them anymore. Seeing a few others starting 5-10 years later and quickly outrunning me is quite discouraging, but such is life.

  > "I spent quite a bit of time on side projects and helping (as I thought) others with projects that I like. After many years of doing this I have come to the conclusion that I simply cannot produce really good code."
I have essentially no hobbies outside of programming, and a ton of side projects + OSS contributions in free time (check my GitHub) and I'm still no x10 dev.

Who cares if it's "GOOD" code -- did you have fun/did you enjoy it? It surely couldn't have been SHIT code, because you cared even the slightest degree.

Statistically, I figure most people writing code don't give a rat's ass and are there to pay rent.

"GOOD" is just skewed by the fact that you can also write code so brilliant it changes the course of humanity -- that's quite an upper bound to place on "great" or "10/10" code.


On the topic of "x10 Developer/Engineer" and terminology:

You know "that 1 genius kid in the grade/school" who is just on a completely different level -- that's been my experience with these "x10" programmers.

There's a 17 year old in the Scala Discord (christ, she's legally a CHILD!) that I thought was a 30-something with some higher degree and was teaching me a bunch of stuff. I about spit out my drink when it casually came up in conversation from some older regular that she was 17.

Half of what she says I can't understand.

That's the sort of thing that comes to mind, for me anyways. Even though I live and breathe code, I think I'd have to be delusional to think I'd ever be as competent as her.

I could maybe memorize the amount of KNOWLEDGE she has currently, but there's a depth of understanding that I can't begin to comprehend.

I’m not sure if you described 10x people or one-track minds.

I love programming, I’m good at it, but what I do with my free time isn’t necessarily related to programming but is still definitely productive. I read up on economics, politics, maybe write a few essays or short stories, work out, pick up a new skill or something. Somehow being a holistic person is now considered mediocre?

Those with one track minds tend to excel specifically at the one track they dedicate themselves to while failing at other things. There are also those who are good at many things. Both have valuable skills (and both are actually well compensated and highly necessary for companies, IME). But if you want to be that guy who can just crank out code for any problem, then I believe this ability is based on the amount of time you've spent coding, which, for those who have one-track minds, is naturally going to be a lot more than the average programmer. It takes all types. There is no value judgement in my descriptions.

I think it is more a case of “talent x perspiration = skill”. Everybody can get better by working at it, but some people advance more quickly.

> You think John Carmack was the only one trying to make 3d games when Doom came out? Thousands of programmers where trying to. Does that make him a 10x programmer in your opinion? Or more like a 1000x?

I think the idea of 10x programmers gets so much pushback because it's often bundled with this kind of toxic hero worship. There's a difference between acknowledging the impressive ability of outliers, and idolizing heroic one-man efforts as the pinnacle of what we should all aspire to as software developers.

Incidentally, Carmack was not the only programmer on early id Software games, and Doom was far from the first 3d first-person game made. Arguably, it wasn't even that ambitious compared to, say, Ultima Underworld, but the programmers working on that get much less immediate recognition.

As a matter of fact, these discussions often remind me of that famous IGN quote of Warren Spector. Except we really should know better on HN.

Warren Spector's quote, for reference:

>"There's a tendency among the press to attribute the creation of a game to a single person," says Warren Spector, creator of Thief and Deus Ex.


Theres also this insane idea that people are just come out of the womb as 10x devs, as though they are literally different breeds and there is no amount of studying or effort that could bridge the gap. Its like a bunch of people who still believe their dad is magical & best and haven't discovered that we're all made of mostly the same stuff.

This is the opposite of how you should want any knowledge worker to think. If you believe that there’s some innate ability that makes people 10x better at a skill than you rather than hours of practice and study then you are much less likely to acquire that skill or learn those skills efficiently. You need to have a growth mindset to be an effective knowledge worker not a fixed mindset.

This is discussed in the book Make It Stick on learning science, in chapter 7 they discuss how your own views on whether skill is innate vs the result of practice affects your own ability to learn and grow.

> This is the opposite of how you should want any knowledge worker to think.

Some of us don't choose our beliefs based on what is most convenient or most beneficial to us. Instead, we try to see the world for what it is, truthfully (or as close to a true representation of the world as it's possible to get, anyway). If having a realistic worldview hinders my ability to learn, then so be it. Not everything is about maximizing productivity.

> If having a realistic worldview hinders my ability to learn, then so be it.

I think the point is that the worldview you described is not realistic.

Many people are successful because of hard work and dedication. Sometimes having talent in a tangential field (design, reading, being a good listener) is enough to make you an excellent engineer or team member. Success is complicated and often depends more on the world than on your particular attributes. Example: Maxwell was a true genius, but he had Faraday's life's work to read through. Faraday was a hard worker and a meticulous note taker who didn't really know math, but without his work Maxwell would not have been able to write his equations. Can you imagine if Faraday gave up science because he wasn't good at math?

One thing I've learned in programming as well as life is that Sources of Truth are deeply tied to individual perspectives.

I feel the core idea is that hard work and deliberate practice will move both those with talent and those without forward and fretting over things you cannot change is an exercise in futility, and you're rather focus on things that you can change.

I fully you agree with you here. But that's not the core idea in the post you're referring to. The core idea in that post was that beliefs should be chosen based on utility, as opposed to truthfulness. I disagree with that idea, because I think truth has inherent value in and of itself. I want to believe things that are true, even if those beliefs are harmful to me.

Furthermore, the poster has very clearly expressed in subsequent posts that they don't believe talent ("innate ability") exists. I strongly disagree with this idea as well, and there is a mountain of evidence to support my position.

If you read the book the researchers argue that worldview is neither factual nor is it helpful.

> factual

Nope. Just because it says so in a book doesn't make it true. When you make a statement of the form "X does not exist", a single counter-example is sufficient to prove you wrong. Lucky for me, many such counter-examples exist, and many have even been discussed in this thread. I'll add one to the list: tourist (Gennady Korotkevich). He's a competitive coder that's several orders of magnitude (e.g. 1000x) better at coding competitions than the average competitor. And the average competitor is a few orders of magnitude better than the average professional programmer who doesn't have algorithm experience. All of this is measurable (at most we can quibble about whether the skill differences are 100x or 10000x or infinite). If you don't believe me, please go and take a simulated Codeforces competition right now and see how long it will take you to solve some set of problems that tourist solved in under 2 hours (or if you will be able to solve them at all - I certainly won't be able to solve the hardest Codeforces div1 problems no matter how much time I put in).

I have a feeling this mountain of empirical evidence was insufficient to change your mind about this. I would appreciate if you could take the time to explain why? Why is this evidence not enough?

I think you’re missing the point or I was unclear. The point is not that masters at a skill don't exist. The point is that attributing their skill mostly to something innate rather than practice, study, etc is not supported by the body of research they review in the book (the book is written for a general audience, their academic research supports the arguments in the book).

There is a mountain of evidence to the contrary. No matter how much you or I would practice, study, etc. we would never become as good as tourist. Most people reach their plateau in competitive coding within 1 year.

For example, here's my Codeforces profile: https://codeforces.com/profile/baobab

Started in 2015, plateau'd in 2016 and no measurable improvement from subsequent 3 years of practice and study.

is it at all possible that practice and study are insufficient to create further growth? What I think faizshah is talking about is that there is enormous complexity that goes into creating skill.

what sort of community do you leverage for their knowledge? how frequently do you expose yourself to new ideas? do you practice abstraction in places other than code? do you practice effectively and take notes that you review? keep a journal? meditate? are you in good cardiovascular shape?

all of these are the tip of the iceberg when it comes to contributing factors. a plateau is a sign of stagnant processes.

I apologize in advance for what I'm sure is a frustrating tone. I'm sympathetic to your situation.

> is it at all possible that practice and study are insufficient to create further growth? What I think faizshah is talking about is that there is enormous complexity that goes into creating skill.

Sure there is enormous complexity and many factors that go into creating skill, but that's a generic claim, whereas faizsah was making some specific claims. In particular, they were claiming that talent does not exist (or "innate ability" is not a significant contributor to creation of skill, if you want to phrase it more softly). There is a mountain of evidence to show that innate ability is one of the largest differentiating factors between "masters" at a skill compared to "normal people" who practice that skill.

> a plateau is a sign of stagnant processes.

Nope, there are limits to achievable skill. If I practice height jumping, I'm going to plateau pretty quickly and make only incremental improvements even with the best of processes. Same goes for mental pursuits like competitive programming as well.

... And the limits are personal, and can vary immensely, unfair as it might be. Some people high jump 2.45 meter, or design a spaceship that flies to the moon, whilst most people couldn't do this in a thousand lifetimes

Usually, though, people don't do something for long enough to get close to their limits, so, generally in life, all this doesn't matter that much, does it.

And to keep trying and doing what one enjoy is good advice :-)

Chess is another example -- two people, the same amount of practice, can yield vastly different results

Thanks for the Codeforces link :-) didn't know about

> we would never become as good as tourist.

Huh, "tourist"? Typo for "Turing"?

No, not a typo. "tourist" is Gennady Korotkevich, I explained this upthread.

Ah, gotcha, found it now. Sorry, I'd missed that, and was thoroughly confused by the pseudonym.

As your sports coach always said "You don't need talent to practice!"

I'm not a huge believer of some mystical "innate ability" that people sometimes use to explain talent. Imagine a great concert pianist who plays flawlessly. After the concert, someone says, "Oh, you were born with such a wonderful gift of talent!" If I were that pianist, I'd be kind of offended--it totally ignores the decades of daily practice it took to get that good. Or a chess grandmaster. You're such a natural! Oh yea? STFU, this took years of grinding!

Talent is used as a label by someone who doesn't understand the practice involved in the given craft.

No one wakes up and paints a masterpiece. It takes years and years of hard work and studying to get to a point before you can even get close to masterpiece level.

But conversely, most people will never produce a masterpiece regardless of how long they spend at it.

The 2 aren’t mutually exclusive. You need talent and hard work to be “the best” in any given field. It doesn’t matter how hard I practice, I’m not going to be a world class sprinter or a world famous singer, my body isn’t built for it.

For some reason when it comes to physical limitations we accept it. But we reject it when it comes to mental capabilities. There are naturally smarter people, at least in specific areas.

People understand things they can see, like, someone is short, then, how good s/he can get at basketball is a bit limited (right). Or, someone having a body shape good for running, but won't lift much weights.

But mental capabilities are invisible, and then it's harder for many people to see that there are different and personal limits.

The capacity to practice is itself partially innate. Take a person with ADHD - they will naturally have a harder time practicing piano. I myself was unable to practice piano as a child, but for different reasons.

> Take a person with ADHD - they will naturally have a harder time practicing piano.

Or easier. Depending on how the hyperfocus-or-none dice roll for piano...

You can be both. I can get better at programming and enjoy beating my personal best but acknowledge some people may be genetically disposed to be better at it even when not trying.

Same reason there is a gaggle of personal trainers at the park near me every morning but none of those working out are thinking they’ll compete in the Olympics

There is an innate ability that drives it to a large extent: inherited intelligence. Recognizing this might not be helpful to the individual's motivation but that doesn't make it untrue. Not everyone can be a 10x (or 100x) Ramanujan-level genius, you really need to win the genetic lottery, in addition to working very hard, for elite level cognitive performance in STEM domains.

Are (were) you a fighterpilot for real? (I think then one needs to do cognitive intelligence tests?)

Couldn't agree more, thanks for this perspective!

When you say "[s]ome programmers will do in 1 day what you wont achieve in a life time" you are no longer talking about the 10X programmer myth. You're talking about invention, rather than day-to-day productivity.

They're the same thing if your day-to-day productivity requires invention, as it does in Carmack's case (and I would argue most programmers)

They are not the same thing, unless they achieve what a 1x programmer won't on a lifetime on a day-to-day basis.

If the 10x, 100x or 1000x programmer does not do the 10x, 100x or 1000x times on average on their day to day then they are not 10x, 100x or 1000x programmers.

(for whatever the heck a 1x programmer is, which people talking about 10x programmers oddly always skip on giving any decent definition)

Yes, exactly.

The best programmers can produce code every day that a mediocre programmer cannot ever produce under any conditions. That is the entire point of the discussion.

I don't even think it's that controversial of a statement, just shift to another field and I think it's self-evident. Paul Erdos could on any day produce better mathematics than the rest of us could produce in a lifetime. Not better maths than his peers, but his peers were also exceptional. Better than the average person who knows how to do their times tables though? Absolutely.

Most people might never realized things like that, because they aren't good enough at math (or software) to see the difference between an ordinary everyday mathematician (if there was something like that -- maybe a high school mats teacher), and Paul Erdos.

Maybe from most people perspective, everyone else is about as bright as everyone else. Because they cannot tell the difference. (And instead people fallback to things like how others dress and how self confident they sound)

I would argue that that's not the case for most programmers, at least not in the sense of developing a new p2p protocol, or a fast inverse square root algorithm. Would John Carmack code, e.g. a checkout system for an iPhone app 1000x faster than most programmers? I doubt that.

The interface he produced for that checkout system, the naming conventions he chose, the nature of state encapsulation, the testability of it? Yes, I believe Carmack would produce _better_ code in all those things than the median programmer every time.

There is much more to 10x than how many lines of code you can produce. A monkey on a keyboard can type faster than Carmack but productivity is not assisted. No one is arguing that a 10x programmer is more efficient by way of swift fingers.

How does the invention happen if not through day to day productivity?

I think the "10x programmer" as a term is confusing.

In one analogy it could be the 10x bricklayer. One who does a well defined job 10 times faster than the others. I think this is the notion people oppose.

Another analogy is the 10x classical composer. A person who produces music that is 10 times better music. But this is much more about creative/artistic capacity than about producing. Nobody oppose this.

I that your examples incarnate the same idea: It is not about producing code faster. It is about having better ideas.

> In one analogy it could be the 10x bricklayer. One who does a well defined job 10 times faster than the others.

Software as brick-laying gets a laugh out of me. If someone uses that analogy, that's a red-flag: they don't really understand software (and tech). It's like offshored bodyshops that just keep throwing more warm bodies at a codebase.

Truth is, a 10x isn't someone who lays bricks 10 times faster. A 10x is someone who thinks differently. Instead of building you a something with load-bearing walls, he'll innovate and use an internal steel frame, paving the way for your building to be taller than anything previously considered possible. [0]

[0] https://en.wikipedia.org/wiki/Home_Insurance_Building

I think this is entirely correct. In this framework we wouldn't call it a 10x programmer, but a 10x thinker.

>> In one analogy it could be the 10x bricklayer. One who does a well defined job 10 times faster than the others.

> Software as brick-laying gets a laugh out of me.

I think that may have been precisely GP's point: The quantitatively-exact-sounding "10x" could possibly be used about an easily-measured well-defined task like brick-laying (which usage could be confirmed by the occurrence of variants like "I'm actually a 12x [or only a 9x] bricklayer"), but not about less well-defined / measurable crafts like programming.

> It's like offshored bodyshops that just keep throwing more warm bodies at a codebase.

Yet they are doing software, aren't they ? So bricklaying has to be part of the "writing software" definition.

But very often what makes an idea "better" is that it produces a result faster. I've done exactly this: spent hours trying to solve a problem to no avail. Then walk away, come back to it and see that a completely different approach will solve it in ten minutes.

I doubt the bricklayer will become more efficient by taking a walk ruminating. But sure, some tasks are like that. Those are the tasks where you have 10x producers.

> Some programmers will do in 1 day what you wont achieve in a life time. And not understanding that makes you a bad programmer by my book simply because this is the foundation of the job.

Most projects out there will not benefit from such a person. If you don't understand this, then in my book it makes you a poor employee. I mean, it's great if this 10x guy can solve some challenging problems that most people cannot, but if that 10x guy cannot show how it helps the business, I hope he's not in my team.

Most businesses are complex and messy. Connecting your code to business value is not easy.

Looking at some of the comments here, people are going out of their way to both defend the existence of 10x programmers, while simultaneously working very hard at not putting any kind of metric on their ability (e.g. "no, it doesn't mean they can produce 10x code or replace 10 people or ... it means <something very hard to measure>"). If you're going to insist on some quality that is hard to quantify, then stop using numbers to describe them. They are geniuses, but not 10x.

In my experience:

1. You cannot replace 10 average programmers with a 10x programmer and hope to achieve the same output. Output, by the way, is not defined by code, but by activities that help the business move forward.

2. While there are some good claimed 10x programmers who work well with others, most of the ones I've encountered do not. In my last job, we had a team of several truly brilliant folks. They could dive deep to the assembly level or work at the system level (scripting, etc). Any bottleneck that arose - they would dig deep and figure it out.

It was the least productive team I've been in. They argued amongst themselves on the pettiest of things that had no impact on the business. When asked why I was leaving the team, I literally told them that I want to be in a team that writes SW. That team's productivity was so low I couldn't reasonably claim they wrote SW. It's not that they didn't write code, but that most of it would not see the light of day because of constant bickering.

A 10x programmer doesn't cost 10 times as much as a regular programmer. 10x programmers are real and drive real business value. The people who insist 10x programmers are a myth simply have not worked with one. Do not conflate brilliant-but-an-asshole or prodigious-output-but-creates-a-mess programmers with 10x programmers.

Yet another comment that uses the phrase 10x but does not provide any way to quantify it.

No one is disputing geniuses exist. Calling them 10x, however, is problematic.

> that uses the phrase 10x but does not provide any way to quantify it

Understand your complaint but "10x" is best interpreted as a figure-of-speech and not an exact mathematical equation. It's just a short & snappy sounding label that's easy-to-say and easy-to-type on the keyboard. (My previous comment about that: https://news.ycombinator.com/item?id=13753178)

We use numerical type phrases without quantitative precision all the time:

- He doubled-down on his opinion. (We don't nitpick and ask how can an opinion be quantitatively measured as 2x?)

- The old editors like vi/emacs is a million times better than IDEs. (We don't nitpick about where the 10^6 quantity improvement comes from.)

- Microsoft decimated the competition. (Some might nitpick that exactly 1/10th didn't get eliminated but Websters Dictionary says that so-called "correction" is wrong anyway: https://www.merriam-webster.com/words-at-play/the-original-d...)

And yesterday, the top comment[1] in "Economics Programming Languages" wrote:

- >a new language to be adopted it needs to do something (something reasonably important) 10x better than the competition. Even 2x better is not enough to motivate the disruption of changing languages.

Yes, it's hard to measure "10x" in languages beyond synthetic benchmarks. But I think most of us get the idea that "10x" is a just a synonym for "a massive amount" of a fuzzy quality.

For some reason, "10x" attracts a lot of extra nitpicking that we don't consistently apply to many other examples ("double-down", "million times", etc).

[1] https://news.ycombinator.com/item?id=28784181

Most of your examples are commonly used phrases. "10x programmer" is a unique phrase. We don't hear "10x electrical engineer" or "10x politician".

Phrases like "decimate", double-down, etc are part of the English language, and is well understood to have multiple meanings.

Your other examples (e.g. vi/emacs being a million times better) are examples of random numbers people throw out. Someone will say vi/emacs is a ton better, or an order of magnitude better, or a million times better. Whereas with 10x, it's always 10x. I don't hear people talk about 3x programmers or 20x programmers.

The HN comment actually supports my point. You can see how different people are interpreting 10x in the responses. That's the exact same problem we have in this comment thread.

>Most of your examples are commonly used phrases.

Yes, and "10x" is itself also becoming a commonly used phrase that's getting less tied to the math number 10. This thread's extensive non-quantitative usage from many people is evidence of that. (I'm guessing that in a few decades, the bikeshedding about "10x" will eventually stop and it will be accepted as a non-numerical description like "a million times better")

>We don't hear "10x electrical engineer" or "10x politician".

The "10x" is relatively new compared to "double-down" but it's spreading out to other uses besides "10x programmer":

- "10x manager" : https://www.google.com/search?q=%2210x+manager%22

- "10x author" : https://www.amazon.com/10X-Author-Level-Left-Behind/dp/10841...

- "10x <various_nouns>" : https://www.hugo.team/10x

- "10x" spreading out to company names, etc : https://www.google.com/search?q=10x

- "10x programming language" : yesterday's HN thread comment example

Why do all the above non-mathematical usages keep happening? Probably because "10x" sounds cool. We're witnessing language evolution as it happens.

>Whereas with 10x, it's always 10x. I don't hear people talk about 3x programmers or 20x programmers.

Sometimes people talk about 100x and 1000x and infinityX to try and emphasize extra rare skills etc. Again, interpreting 1000x literally as 10^3 isn't the intended meaning. And to build on your point... the fact that nobody says "9x programmer" or "11x programmer" but almost always "10x" is actually evidence that it's not trying to communicate exact mathematics.

> This thread's extensive non-quantitative usage from many people is evidence of that.

Even putting aside comments like mine complaining about the actual number, this thread's extensive use of 10x to mean as many different things is a good reason to avoid the term altogether. The whole thread is just "10x programmers don't exist because X" and "10x programmers exist because Y", where X and Y are unrelated.

> "10x manager"

All but one of the links in the first page were in the context of SW, or written by people in the SW industry. Many of them explicitly pointed out they derived the phrase from 10x developer.

I cannot count this as an independent use of 10x.

> "10x author"

Written by a tech guy.

> 10x <various_nouns>" : https://www.hugo.team/10x

Written by a guy who started a SW company.

> "10x" spreading out to company names, etc : https://www.google.com/search?q=10x

Sad, but fair enough.

It sounds like 10x is the equivalent of putting 2.0 on everything (and also as meaningless).

>10x to mean as many different things is a good reason to avoid the term altogether. The whole thread is just "10x programmers don't exist because X" and "10x programmers exist because Y", where X and Y are unrelated.

Yes, I understand your complaint here too but the various X-Y meanings isn't really the fault of "10x"... it's caused by any label. Previous comment about that: https://news.ycombinator.com/item?id=28797871

E.g. the alternative word "expert" or "effective" such as "effective programmer" would cause the same debates:

- I think an effective programmer is one who understand the whole stack from hardware gates to web stack

- No I think a truly effective programmer is one who empowers his team members.

- No an "effective programmer" is really X. No it's Y.

- <... ad infinitum disagreements ... >

It doesn't matter what the word is... "10x", "talented", "expert", "master", etc. There's no consensus definition and yet we haven't tried to eliminate those words.

Generally, I understand that people typically mean "10x" as a synonym for "massively better". (Because nobody who says "10x" has a stopwatch and rigorous academic studies measuring it.) And yes, the counterargument is "10 doesn't really mean anything" ... that's true but the "'massively better'" also doesn't really mean anything -- and yet we can't strike "massively better" from our language so we're back to the same issue.

There is no short label X that "really means" whatever everybody agrees it to mean. That's human language. We muddle onward regardless.

Fyi ... a correction as I've since listened to the book author's background since this thread. (https://www.youtube.com/results?search_query=sean+platt+inte...)

>> "10x author"

>Written by a tech guy.

The author of that book, Sean M. Platt, is not a tech guy nor a programmer. He's a high-school dropout that started writing articles and stories. Now, he's mostly a publisher.

The interesting thing is that somehow, the "10x programmer" meme made its way to a non-programmer and he adopted it as "10x author".

> They argued amongst themselves on the pettiest of things that had no impact on the business.

> It sounds like 10x is the equivalent of putting 2.0 on everything (and also as meaningless)

What do you mean by that ? If I write a cooking book titled, "pizza 2.0" everyone in the world will understand what the 2.0 means.

It's 2 the latest version of the book?

Or a new better pizza

The problem is value and productivity in software (or any creative space) is hard to quantify in general. How much better is the best engineer you’ve worked with compared to the worst? 2x? 1.5x? How do you quantify that? But you know one is significantly better than the other I assume.

If you want to replace 10x with “much much better”, then ok. But the general idea is the same. Some engineers produce an outsized impact compared to their peers.

How is it a problem?

Carmack is an interesting choice here, oddly. He and team were among the first ones there, but it isn't like they weren't rapidly met. Nor, strictly, the first. Ultima Underworld, I'm pretty sure, has a bit of that distinction.

Even more impressive, much of the success of later titles came from his picking the likes of Abrash to help make things fast. Notably, the WTF comment in the "fast square root" trick was from Carmack about a contractor's code that was speeding things up.

Getting even more meta on it, they had amazing dev tools at their disposal that many other programmers just couldn't afford. Look at the recreations folks do nowadays using pico8 and friends. Some folks get some really amazing tricks running in that field.

None of which is to say Carmack is a bad dev. Quite the contrary, he is quite good. But don't discount the role of everything else to the early success that Doom and friends had. In particular, the artistic execution and vision of everyone else was probably worth more than you'd consider.

I don't know about producing in one day what it would take another person 2 weeks to accomplish, but I do know there are some core features of my company's app that were created by a very talented developer that I don't think could be reproduced by some of our junior developers in a year.

I think it's a mistake to conceive of the difference as one between the tortoise and the hare, where the tortoise gets there, eventually. There are bits of this code that the tortoise just wouldn't ever end up producing if you gave them all the time in the world.

The myth part is that you can hire these 10x programmers by asking harder whiteboard coding questions and they will sit and close jira tickets for your company thereafter.

I think that depends on the context in which you’re working. Flashes of genius don’t have much place an an enterprise code base. Everything is rigorously standardized and separated into layers. Each layer has a bunch of bookkeeping about dependencies, interfaces, mocks, expectations, and test cases. This is not really brain work; it’s typing. Now maybe if you’re really brilliant, you’ll do much better than average with the implementation of a really tricky layer. But the layers are rarely tricky, implementing them never takes all that much time, the slow part is always the integration and associated mocking rituals. Typing speed pays better dividends than brainpower in terms of productivity, there.

Maybe that's why so many people think 10x programmers don't exist - they work in environments where a 10x programmer isn't needed and couldn't really demonstrate their skills.

How many people will never achieve their potential because they don't work in a room with a door?

So the 10x programmer realizes you can accomplish the same goals with 1/10 of the book keeping, dependencies, interfaces etc. etc. etc.

In other words, the processes themselves place hard upper limits on the possible productivity.

One could make the argument that skipping the bureaucracy catches up with you in the long run.

Or maybe a small team of 10x programmers could really get away with it, indefinitely. But can you reliably staff enough of them, now and forever, to keep that going? For a business it is often better to have a codebase that scales with arbitrarily many average-quality engineers, than to be dependent on hiring miracles. Even small numbers of them.

I find this true to an extent, but this also where a lot of unnecessary complexity, risk aversion and process live.

This is always my go to example of a 10x programmer:


How many developers do you know that would bang out a spelling corrector on a single plane flight, in half a page of code?

It's a question of knowing the algorithms of tools that make things possible that other developers might not even consider, and how to express those ideas cleanly and concisely, evaluate your solution, and iterate and improve quickly.

And then consider the quality of his documentation and considerations for improvement and future work.

Norvig wasn't given a random problem that he then solved quickly. He picked the problem to work on. I'm sure I can give him problems of similar complexity that I work on for my job and he wouldn't do it that quickly. Quicker than me, for sure - but it wouldn't be hard to find someone who does it quicker than he did.

I'm sure I can find problems that I can solve quickly and amaze people. Being able to pick your own problems makes for a bad metric for identifying 10x programmers.

This is one example.

Norvig has written a ton of kick ass programs in his career that are concise, efficient, and correct, in much less time than most programmers would require.

Furthermore, even picking your own problems can be a way to achieve > 10x productivity, as identifying useful problems to solve is also a crucial skill.

I attended a talk given by Norvig, where he frankly admitted Google picked markets to enter where they could get good quality data to train their machine learning models on (this was many years ago). Maybe lesser developers wouldn't have that insight.

People can be 10x in one domain but not in others, experience is part of the equation. That's not too strange is it? Compare a football player in a ice hockey team. Even though they still have a significant head start compared to someone on the street, when it comes to physique, ball sense, etc.

> It's a question of knowing

Exactly. Norvig's skill set happens to be of the upscale variety, but similar "10x" knowledge+practice effects exist in more pedestrian domains as well.

Compared to my current self, my teenage self was a 10x PHP programmer. I'm absolutely sure that it would take years for me to get back to being even remotely good at PHP-interleaved-with-HTML style coding.

I think the fundamental issue with 10x is that a lot of people simply never met one. These guys tend to fly under the radar quite a bit. If you are in a tier 2 market or company, your chances of attracting one are close to nil. Because they are extremely valuable, they don't interview a lot and tend to hop between companies where they know people (or get fast tracked internally).

I remember someone laughing at the idea of 10x programmers. I asked him where he hired from and what's the compensation like. Turns out two high performers I had recently hired were from the same college he was trying to hire from. Of course, he had no chances of attracting anyone since my offer was ~3x what he was offering!

Author nails it though with: "We should be far more concerned with keeping 0.1x programmers off our teams than finding the mythical 10x programmer."

No, you need to watch out for the -1x programmers. At least the 0.1x are moving forward,

I agree, because the effort required to filter out a 0.1x programmer is vastly smaller than trying to find a 10x outlier. Way more bang for the buck.

This is just rescaling. The author is actually acknowledging that there exists 10x differences but divides by 10 for some reason.

The author's point is that more effort should be spent removing negative outliers instead of hunting for elusive positive outliers.

1x is supposed to be a baseline for an average productive programmer.

0.1 is someone who isn't actually a productive engineer but somehow faked the interview process.

It's not difficult to be 10 times more productive than somebody who isn't productive at all. I'm at least 100x better Kotlin developer than my 30 years dead grand-grand father who never saw computer IRL.

It's much more difficult to show that you can be 10 times more productive than the average productive developer.

You don't say that so-and-so is a "10x painter". Art is obviously not about the quantity of paint used per day, but how it gets used and in what context. This is what you seem to be communicating with your quote about bittorrent.

Because we're programmers, we chose an empirical measurement ("10x") as our term for high skill. This terminology is confusing and stupid, but now we're stuck with it

I'm guessing there are clearly 10x painters other painters can identify, but there are even fewer quantifiable metrics to measure a painter's output than there are for programmers.

> You know, 10x is an optimistic number here. Some programmers will do in 1 day what you wont achieve in a life time

Exactly. There are many modes of that, and you named a few examples. Another example, company has 2 systems maintained by 10 people each. Someone comes in and figures out how to generalize one of the systems to both cases, so they free up 10 of those programmers to go do something else.

Or - 10 people are beating their head on a problem for months, someone comes along and says "it might be OK if we don't solve that problem, let's move on"

Or - someone figures out that a system/module doesn't need to be written because an out-of-the-box solution exists.

Like, I am not a 10X programmer in the sense that I can write 10X the code of the next guy. But have I prevented 10-people's-worth of wasted effort in my career? Yeah...

If nobody had previously succeeded then I would take the point that Carmack was somehow unique, but by the time Doom came out lots of programmers had produced 3d games, and we had ray-tracing and lots of graphics theory and implementation (e.g. elite written in 6502 assembly on an 8 bit cpu).

>You think John Carmack was the only one trying to make 3d games when Doom came out? Thousands of programmers where trying to.

Yes Doom was a great game, and well programmed, but it didn't really move the state of the art forward massively, as I would argue that hundreds of programmers had already written and published 3d games by that point.

What do people gain by arguing over these? Genuinely confused

Just to be proud and gain that dopamine?

> What do people gain by arguing over these? Genuinely confused

If you assume some arguments (eventually, in some way) lead to people altering their opinions closer to the truth, then people gain more accurate views of the world. This also reduces the spread of misinformation (when the people who have corrected their views no longer spread it), thus having a compounding effect on more people getting closer to the truth.

Or if you assume people never change their opinions as the result of arguments, then I suppose people gain nothing by arguing.

@VMtest, Personally I think these discussions are interesting as a small part of learning and reading about hiring and recruiting.


> if you assume people never change their opinions as the result of arguments, then I suppose people gain nothing by arguing

One thing: they can slightly learn how others can react to what they say, and thereafter, in real life, better know what to say, or not.

I'm about 1/1,000,000th a Fabrice Bellard, and I'm fine with that.

Bellard is who I reference in these situations also. He's a sort-of empirical refutation of the assertion "10x programmers don't exist". I always liked that he lists qemu and ffmpeg on his website as afterthoughts mixed in with all the other code he's written.

Bellard is obviously 10x (or 100x) engineer when you need to port Linux kernel to JavaScript or compute the longest Pi sequence.

But most jobs aren't like that and require a mixed skillset - problem solving, people skills, stress resistance etc. and it's entirely unclear how Bellard would fare there.

But the question isn't about those skills, the question is about programming. Bellard is blatantly an insanely productive programmer.

Also if you think a man who can build an LTE base station from scratch isn't good at "problem solving" I really don't know what metric we're talking about anymore.

No, this is about being 10x engineer and part of the job description is much more than programming.

On the edge of possible/impossible, ratios will lose their meanings. In the same way, you can say "Einstein is a 10,000x scientist."

Continuing with this analogy and using Kuhn's concepts, maybe it's better to bracket out breakthroughs (paradigm shifts) when talking about everyday productivity (normal science).

The point is 10x what? Compared to who?

We all know people who code faster than the rest of the team but they always cut corners. Part of the reason that everyone else is slower is because they're trying to work around the under-tested tightly coupled mess left behind by the 10x guy.

Then there's the 10x complexity discussed in the parent comment. Plenty of people will take years to solve a problem, if at all, that someone else would manage in a month. However this may be reversed for a different problem since it depends on experience and predisposition, so it makes no sense to talk about a '10x engineer' in this regard.

I share the conclusion with the OP that it's a vague and ultimately useless distinction.

Mess creators that slow the company down isn't what's meant with 10x.

What you're describing sounds more like a -0.5 developer

The confusion is that it's all about what people can and cannot do, not how much code they can write.

If your development needs are mundane crud stuff, you don't need a "10x programmer". If you're doing tricky stuff, on the other hand, you absolutely do.

I do CRUD stuff mostly at work but I always research things to improve. A good programmer can make a huge difference in CRUD applications. Things like proper relational database design can save a lot of issues down the road. Using good OOP practices can improve reliability. Keeping things DRY is great too.

I have seen some awful source code that run business applications for years.......but bad engineering practices may cost millions in the long run one way or another.

I shouldn't have said CRUD. I agree with you, there's a place for genius stuff everywhere.

What I meant is that there is also a lot of boring stuff that just needs to get done. A 10x programmer won't help you much there.

Even if you are doing mundane crud stuff, you are eventually going to run into hard problems that don't have obvious answers. Not everyone has to be able to solve those problems, but it's useful to have someone in the team who can.

I experienced people who are 10x better than me through Google codejam.

I looked at one of the problems of the final round, looked at the solution, read it for hours, and I couldn't implement it correctly.

Then, there are these top 3 who came up with their own solution and wrote correct code within 20 minutes.

For comparison, I can consistently be in the top 1000 of Google codejam. So, I consider myself a fairly good coder.

I experienced this with topcoder as well. There are some people who are way too smart.

10x programmers and engineers are very much real. It is just rare.

what would be nice is turning the skill gap into motivation to make other people grow faster

but the reflex for humans is to fear and be angry (understandably, but a bad reflex)

This is a very insightful list. But there is one thing I'd quibble with: while it's inherently a super-subjective metric, I would say that 10x programmers do exist. Both in the "can support a company by themselves" sense and the "mad lone genius" sense. Not all the 10x programmers are good at working with other people or on other people's ideas.

0.1x programmers are often lost or afraid, although they may just be unimaginative. The real risk of fear is introducing 0.1x processes. People are often willing to sacrifice any amount of development velocity to avoid being seen to make mistakes.

Sometimes this makes sense, sometimes it doesn't. Arguably the reason for the success of SpaceX is that they were willing to fly, crashland, and improve a dozen rockets in the time that a conventional company would spend agonizing over whether to launch at all. On the other hand, nobody wants to be on the same roads as the "move fast and break things" self-driving car.

Saying 10x programmers are a myth because it's crazy that someone could do in 1 day what another could do in 1 week makes the assumption that both programmers in the example have the same amount of context, background understanding, and experience.

The benefit of talented engineers comes from the fact that they avoid costly pitfalls - not that they pump out more code per hour worked.

I feel this "10x programmers are a myth" attitude comes from a good place of wanting to promote harmonious working and skill-up junior engineers, but it seems so totally detached from reality. I'm not sure the "benefit" of propagandizing against talented assholes is worth the cost of demotivating people who might be inspired that they could become a 10x engineer.

Realistically, I’ve always found it a weird sentiment more because issues caused by software developers are not of the “this person is 1/10 as productive as this other person” type. A bad developer isn’t a low positive effect on a project, but an active negative effect. As such, you can put literally any Y for a (Y)X programmer, since the scale of how inefficient a programmer can be goes up to infinite.

I think the pushback against 10X programmers comes primarily from the fact that many programmers who look good to managers (constantly push out new features) are in actuality an albatross around the necks of the other programmers who constantly have to fix the messes they create (in reality, a negative on team-wide production). What matters more to me in a teammate than how productive they are is to ensure that they are on the positive side of the scale at all, which the perceived 10X programmer is often actually a failure at.

Though I’ve definitely met people who just have such a wide knowledge base and think so quickly that they are in fact 10X as effective as a standard productive programmer, so I do understand where the idea comes from.

More people should be aware of Comrade Stakhanov, the "10x miner": https://en.wikipedia.org/wiki/Alexey_Stakhanov

> In 1988, the Soviet newspaper Komsomolskaya Pravda claimed that the widely cited achievements of Stakhanov were puffery. The paper insisted that Stakhanov had used a number of helpers on support works, while the throughout was tallied for him alone. Still, according to the newspaper, Stakhanov's approach had eventually led to increased productivity by means of a better work organization, including specialization and task sequencing

Exactly the same discussion, except there it was tonnes of coal which can be easily and unambiguously measured. But they can't be easily attributed. How much of his work was really that of his team? And how much of it was simply inflation by his managers for propaganda purposes? And how much of it was, underneath all the propaganda, real process improvements?

TAN: I fearst heard of "Stachanovism" over forty years ago, in my teens. Some time before that I had read an anthology of American folk tales, so to me Comrade Alexei Grigorijevitj was always kind of a Soviet version of John Henry. Made it easier, when I later found out, to accept that Stachanov was also at least in part a myth...

Yes attribution can be hard.

At the same time, if one compares shoveling coal and laying bricks with software, then it'd be good to smash and destroy bricks, throw the coal back into the mine, raze houses. (Reducing the amount of code)

Software would be more comparable to designing the coal mining site, or architecturing a city (but not building it). And then maybe it's simpler to see, that individuals sometimes can have much impact.

> The benefit of talented engineers comes from the fact that they avoid costly pitfalls - not that they pump out more code per hour worked.

Not just avoid pitfalls, being able to put together a good architecture is a huge factor as well. The effect is cumulative. If you have a good design to begin with, everything becomes easier with less risk of bugs. The result is that implementing new, or changing existing functionality takes less time.

The 10x developer doesn't write 10x as much code, (s)he will probably write less code than the 1x developer.

Exactly. It's not how fast someone churns out code but how well thought out/elegance in approach that delivers this 10x value.

Instead of writing entire systems from scratch, maybe you have enough experience or familiarity with well known techniques and libraries to ship faster.

> (s)he will probably write less code than the 1x developer.

Or better yet, remove code.

This needs to be required reading for every first time, non-technical manager of software developers.

I always read it as a direct comparison though.

It doesn’t matter if I have better experience or I can contextualise intent better; the phrase “10x” means it would take a team of 10 to do it, in the time, scope, budget (whatever).

But there are major reasons this could be the case, if I’m on my own and I know the direction then I don’t have to quibble about what framework to use, how to name my variables, what I consider “proof of concept quality”. You just get your head down and do it with the entire scope and progress in your head at once.

No infighting about semantics, no deeply technical progress reports or trying to subdivide tasks which are not inherently subdividable.

More people is more communication and it’s quadratic.

The author exactly says this "two engineers with the same experience".

I'm sure there are engineers that can create an architecture 10x better than mine.

But show me someone that can create UI 10x faster and I'll admit they're a God.

Some things can't be 10xed after some point. It just takes work.

Some things can be 10xed and some just can't. I find it silly to think someone can be overall 10xer at everything.

> Saying 10x programmers are a myth because it's crazy that someone could do in 1 day what another could do in 1 week makes the assumption that both programmers in the example have the same amount of context, background understanding, and experience.

Yes, because that's how it's always presented. And if it's not, then it's also a myth because then it's not a 10x dev, it's 10x context, background understanding, and experience.

Either way it doesn't hold water.

> it's 10x context, background understanding, and experience.

A person who is able to build 10x context, background and understanding is a 10x developer. Most people lack the ability to do that. Your post makes it seem like you could give those traits to anyone easily, but it doesn't work that way building the right mental models and being able to generalize your knowledge are key factors in talent.

Why do people need to be motivated in such a way? Won't that just lead to a lot of disappointment and misery later on? Why isn't "being a good programmer" something we should admire?

Of course one programmer can create 10X the code of another. The limit isn't how fast they type. Its how correct the code is, and how much it leverages what you have and what you'll need. And that results in 'pumping out' more code per day. Whatever the reason (context, background, experience)

> The limit isn't how fast they type. Its how correct the code is

And how good your autocomplete tools are, or you repertoire of code based to copy from.

> I feel this "10x programmers are a myth" attitude comes from a good place of wanting to promote harmonious working and skill-up junior engineers

When comparing juniors you observe the same distribution.

I've met 10x straight out of college who ran around other new grads.

1 day vs 1 week: wouldn't that make them a 5X engineer? That's half of the advertised 10X...

>, I would say that 10x programmers do exist.

Yes but I think the underlying issue is that some people really Really don't like to acknowledge (or are unaware) there's a small set of elite programmers that can do things average coders can't. But any label we use to describe them (i.e. "10x", "rockstar", "ninja", etc) will be psychologically distasteful. I previously commented about that effect:

- https://news.ycombinator.com/item?id=13753178

- https://news.ycombinator.com/item?id=24753594

To see that hidden psychology driving the narrative... look at how the author constructed his comparison with extra qualifiers of "competent" & "similarly experienced" : ">The idea that someone can produce in 1 day what another competent, hard working, similarly experienced programmer can produce in 2 weeks is silly."

In other words, if we artificially construct the comparison by a priori assuming both programmers are equal, then the 10x difference is a myth. Well yes, that's tautology. If you invent a sentence to "prove" a point, you can set any conditions in the artificial universe in your favor that seems to support it.

However in the real world, the stack of resumes will have a candidates with wildly different skill levels. Some struggle with Fizz Buzz. Some can write a 3d game from scratch. But it's impossible to create a label to describe that wide gap which also doesn't invite bikeshedding on that label.

EDIT reply to: >"Why can't we call them master programmers?"

Because we'd still have endless bikeshedding of the label "master" as in "Master programmers are a myth..."

We don't call Vermeer a 10x painter, Beethoven a 10x composer, nor Shakespeare a 10x author. It would be silly to call Turing a 10x computer scientist, or rockstar and ninja would be insulting.

Vermeer, Beethoven and Shakespeare are masters of their crafts. The same word we use for skilled tradesmen. Why can't we call them master programmers?

People typically call them geniuses, not masters. And the reason we don't call them genius programmers is that people don't like to use the genius label for people who aren't famous.

Because we're talking about different things.

A master craftsman will solve a problem quickly, efficiently, and elegantly, combining insightful attention to detail with speed and effectiveness.

A master creator will produce work of lasting relevance and power. The metric is quality and impact, not speed and efficiency.

Occasionally you get people like Mozart who combine elements of both, but they're exceptionally rare. And they work best in a relatively limited context which they have mastered completely. They're not perpetually chasing the latest shiny. (Mozart basically knew one style. Even he would struggle to master all the languages and genres that are common today.)

Turing would likely have made a poor backend developer, but he produced work that had lasting impact.

A typical 10X developer will be more like a master builder than a master computer scientist.

I don't find the labelling offensive, because it's clearly realistic. Some people are just very, very good. They produce clean, tight, code at speed, far more quickly than muddlers who produce reams of mediocre code which creaks along, breaks when you look at it, and doesn't actually solve the problem.

But a 10Xer isn't going to be good at everything. If they mostly do backend, they're not going to go toe to toe with game devs. Etc.

I think part of the reasoning is in the trades there's the presumption that to be a master you must also be old (with exceptions, Da Vinci was active when he was still younger, 20).

In the programming industry, there's likely a bias more toward younger people, and hiring offices ask for the world. They'd like a whole staff of masters and rockstars. Though that doesn't preclude "older" people from being known as masters either. Take John Carmack or Jonathan Blow, for example.

That is a interesting thought, that seems to imply that those young programmers will not improve. If you are a master at 25, what do you call yourself after you have gained twenty more years of experience? It may be that those 25 year old's a more like journeymen then masters, or that programming is more like sports where youthful vigor are more important than experience.

John Carmack programmed Wolfenstein at 21 and doom when he was 22. At that point he did what most "masters of that craft" couldn't do. What would you call him? And what would you call him now, 30 years later?

> And what would you call him now, 30 years later?

Something else than thirty years ago, because presumably he has developed since then.

I think that was why GP suggested "journeyman" for younger people, because otherwise there's nothing to differentiate young "masters" from masters honed by thirty years of additional experience -- except adding the whole phrase "with thirty years of experience", which is pretty clumsy.

How many times faster is an O(n) algorithm over a O(n^2) squared algorithm?

How many times faster is a developer who can write a compiler over one who can not?

We are at least consistent; I rather frequently see programmers claiming a "105 times faster!" benchmark that was actually the result of a fundamental complexity change.

Not everything is on a linear scale.

I once worked on a team that in hindsight had two 10x developers in a team of only 12. They were vastly different, and the difference to me is extremely interesting.

The first one was not a bad person by any means, but had no social skills, no ability to communicate and often just forged ahead building great stuff on his own. Entire areas of the software were entirely conceived and created by him, and it was difficult to even grasp what he was doing while pair programming with him. In a day he would write more tests and get more coverage than the rest of us would create in a week.

The second one had come from a FAANG-ish place where he was extremely successful, but didn't want the money or stress. He was extremely kind, considerate and great at communicating. Pair programming with him was a joy and he helped everyone else on the team grow. I'm confident he could have created entire regions of the software on his own, but he knew that wasn't the right thing to do. He would often ask questions until someone else on the team could solve a problem, even though I'm sure he knew how to solve it easily himself.

Yeah the first one is a risk. I've met one whom I had to take over the code of when he left and psh, I wished he was a dumbass instead. He could produce shitloads (but would never test, interestingly), do stuff I honestly can't do, but do you really need to bitshift instead of doing a modulo to save 1 CPU instruction ?

The amount of obscure waste of programming complexity made us actually just abandon the thing he spent years on to just not do it. And everything works fine and we still make money and we're just probably a little bit less clever about how we do it. We also have a lot more time to listen to users and solve their problems, rather than gloat on our genius and solve our own created problems :D

Your second example is inspiring and I hope to meet one to learn from too !

My go-to examples of 10x developers are benevolent dictators like antirez (of Redis), evanw (of esbuild + Figma CTO), youyuxi (of Vue), Carmack, etc.

Both qualitatively and quantitatively, their output is at least 10x of a typical developer at a typical company, not even counting everything these people do besides writing code. Many more are not famous but are in the same league of impact/productivity - I'm genuinely puzzled how someone can say such people do not exist.

I agree. Another example, Linus T wrote git from scratch to hosting the Linux kernel in 2.5 months. I'd say 10x is a low estimate.

I've worked with a guy who wrote great code like he was typing an email.

Everyone has a ceiling of the complexity of problems they can handle, some programmers are outliers in this regard.

Bit of an aside, but writing an email is its own art. Not many developers know how to write an email that a busy, non-technical VP will read, understand and act on in the way they intended. (Then again, neither do many managers.)

You mix popularity with 10x. Not all 10x programmers are popular.

I believe 10xers exist within a business or technical context. This isn't limited to the field of computer science, but even traditional engineering (Edison) and sciences (Newton) have plenty 10x examples.

I'm not mixing - as I wrote, many more are not famous, but their impact is similar. When discussing abstract concepts, it helps to have concrete examples of that phenomena. Using "famous" people is a shortcut to define a shared definition for what "10x" might look like.

And yes, there are "10x" teachers, investors, lawyers, storytellers, sales folks, etc. Similarly, the distribution of talent that helps with "software development" might have a very long tail, leading to extraordinary performance versus the median.

IF 10x exist, some will be popular and most won’t. Obviously he won’t know the obscure ones or pick examples you can’t relate to. So while he didn’t prove his point perhaps, perhaps you didn’t disprove his either. :)

To pose it another way: for general software Dev (not some niche use case that only one guy understands), let’s say building websites in Vue or React, the idea of a 10x developer is that they can work in January and another, average, programmer can then spend February, March, April, May, June, July, August, and Sept… and the average guy won’t get more done?

Or they come in only every other Monday and contribute as much as a median Dev working full time?

It just doesn’t seem reasonable at all, at least if we read it literally.

I have worked with and met great, outstanding, developers who are much better than me. Some of them famous. None of them are anything close to 10x as productive, and I’m not particularly great. They’re much smarter than me, for sure, but they’re not 10x as productive.

How long would it have taken you or me to write Doom, or Quake? How long would it have taken Carmack, if it'd just been him?

I think you have a false dichotomy here. Carmack isn't 10x better a C or assembly than me. But in his domain he's easily 10x smarter and more productive than me.

And on top of that, he has a quarter century more experience in graphics programming than me. That shit accumulates.

Yes, actually. In graphics programming, from 2D and software rendered, to OpenGL and VR, you could give him and me (and presumably you) any task achievable in a few human decades, and he'd do it more than 10x faster.

"Graphics programming" is not a small niche, and yes in that space he's a 10x programmer.

I think you're thinking of this the wrong way. I agree that for something as simple as "build some websites", yeah, 10x is going to be very hard. You're then just turning specs into code, pretty much.

10x programmers are those that don't waste 9 months. The hard part is knowing which 9 months out of 10, which 9 ideas and components out of 10, is waste.

For "build a web forum" or something, yeah there are no 10x programmers. Though there are many many 0.1x programmers still.

> How long would it have taken you or me to write Doom, or Quake? How long would it have taken Carmack, if it'd just been him?

I'm not a games programmer. I doubt a median games programmer would take 10x the length of time that it would take any other games programmer. It just seems implausible, and I don't know why people want to defend it.

I spent 8 years at EA and was a senior software engineer. I have absolutely witnessed 10x game programmers.

I think in any field where you have a deep domain with complex, difficult constraints, it is entirely possible to have programmers with 10x the effectiveness of the median simply because of their greater domain expertise. They know the unique algorithms, data structures, design patterns, hacks, tricks, hardware quirks, tools, debuggers, code smells, etc. All of that has a huge compounding factor when it comes to output.

Here's a story I like to tell:

I joined EA shortly after the PS2 came out and worked at the studio that made Madden. Madden was a launch title on the PS2 (meaning it shipped right as the PS2 hardware did) and that version essentially sealed Madden's success. It was hugely successful.

But making a launch title is really hard. You're developing softare for a hardware platform that is itself in flux. When EA first started working on Madden PS2, they didn't have any dev kits to work on. I don't know if dev kits even existed. They needed to write code that they couldn't compile and run. But if they didn't start immediately, they would never be able to get the game done in time for the launch.

They had a guy on the team with a reputation for being a wizard. He took the MIPS R5900 reference manual, and disappeared into an office with blacked out windows for some number of weeks. When he re-emerged, he had written a PS2 emulator despite never having access to an actual PS2. The rest of the team were able to compile and run the game on that emulator so that they could make progress until eventually real dev kits arrived. From what I heard, when the real hardware showed up, the game actually ran on it.

Now, some of this legend may have grown in the retelling (it was already legendary by the time it was told to me), but I worked with that engineer a few times and I can vouch for his incredible ability to get stuff done. I remember once when FIFA was having performance problems and couldn't get their framerate high enough to ship. He was called into to help and a few days later, got it up to a solid 30 FPS. This was even though the team's own engineers had been trying to make progress and he had never touched the code before.

I think people dislike the mythos of 10x engineers because it's interpreted as some sort of cognitive essentialism. But I think most of it is just really deep domain expertise. Instead of talking about "10x engineers", we should talk about 10x at what.

The median games programmer wouldn't be able to do it in 50 years. (assuming they were locked in a room for 50 years with that same hardware, not that they got to see state of the art progress and upgrade their machine)

The amount of innovation, from applied BSP, Zbuffer tradeoffs, light maps with interpolation, interleaved FPU corrections for texture mapping, multiplayer protocols (qtest used TCP), brush-levels, etc...

Before ID software did these things, they weren't checkboxes for what to do. After, sure. But the hard part of programming is not typing.

The median programmer would not have been able to do it before it had been completely irrelevant. The median games programmer could do it in 2005. In 1993-1996? Not a chance.

The main competitor to Quake was Duke Nukem, which technologically was just "doom and a half". The enemies were sprites, ffs.

ID software was years ahead of every other games developer.

The reason other games licensed the Quake engine wasn't that it was cheaper than doing it yourself, it was that they couldn't do it.

Edit: granted, I have a low opinion of the median programmer. But sometimes in life you meet people who, with things you think you're actually pretty good at, just completely mop the floor with you without even trying (or showing off). It's humbling.

> I doubt a median games programmer would take 10x the length of time that it would take any other games programmer. It just seems implausible, and I don't know why people want to defend it.

Then why did all those thousands of other game programmers fail at creating a decent 3d game, before Doom was released? It's not like Carmack was the only one trying.

> How long would it have taken you or me to write Doom, or Quake?

Almost definitely not 10 years.

I think it's more likely that we wouldn't have thought something like Doom to be possible until Carmack came along and squeezed oodles of math tricks and hacks into the game to let it run on hardware of the era. So it may as well have be an infinitely long development time for us.

If you wanted to write Doom, starting in 1993, and take 10 years, your most efficient strategy would probably be to do something else for the first nine of them and wait for the technology to change around you.

Waiting for the equivalent of Carmack to write and open source the underlying libraries, basically.

Almost certainly you would just plain have not been able to do it at all. Not in 1993-1996. Not without it already having been done once.

And if you had, it would have taken you more than 10 years to realize what the right thing is, and to solve all very complex problems that showed up on the way.

Quake doesn't even have any load times! Not even with the hardware available at the time.

There is no "10X" touch-typist because a) the performance limit is absolute & relatively low and b) the learning curve is such that it's easy to get to the limit and plateau. Now let's increase complexity - still typing, but you must translate the text from another language. Increase again - now you're translating poetry. Increase again - the poetry is original Shakespeare. You need to translate into contemporary English approachable to Gen Z, as a script for a TikTok short, referencing memes from current events. There is no doubt that some people will have a combination of talents - in language, typing, copywriting, built-up cultural knowledge, lived experiences - that will make them 10X in such jobs.

The more software development work approaches touch-typing, the less difference between 1x and 10X — strict Jira tickets in, garbage out, maybe slightly faster. Spec-work, and most grunt-work, is undifferentiated and does not benefit from a 10X person. The more your work approaches poetry - such as exercising user empathy, improving the architecture, considering UX, respecting performance constraints, understanding user's jobs to be done, and improving your implementation around that - the more creative space there is for 10X to reveal itself.

>There is no "10X" touch-typist

Ironically enough your example is flawed: there are documented 10x touch-typists.

The realities of the court system place high demands on the typists; the minimum required typing speed is already quite high: trained court reporter or closed captioner must write speeds of approximately 180, 200, and 225 words per minute (wpm) at very high accuracy in the categories of literary, jury charge, and testimony, respectively[1] - and some exceed the minimum and go for 300 wpm. Even better, the official record for American English [is] 375 wpm [2].

Compare that to the average typing speeds around 30 - 40 wpm [3] Granted, the court reporters use specialized input devices (stenotypes) - but hey, the same can be said about highly productive programmers, who use specialized development environments - and sometimes also specialized input devices.


[1] https://en.wikipedia.org/wiki/Stenotype

[2] ibid.

[3] https://en.wikipedia.org/wiki/Touch_typing#Speed

You make a good point that hints at another often-undefined assumption in these discussions - 10X performance vs who, exactly? IMO, it should be vs "a typical peer in that profession". For court reporters, if "a typical peer" is a trained court reporter who averages ~200 wpm, and a superstar is at 300wpm, then it's a 1.5X gain. If a "typical peer" is a regular Joe, then 300wpm would be 10X - but this doesn't feel appropriate because we compare productivity of a "10x developer" vs a "typical developer", rather than vs "an average Facebook visitor".

Fair enough.

I've had good luck discussing the subject back in 2013: https://news.ycombinator.com/item?id=6464807 and https://news.ycombinator.com/item?id=6465175, with the insightful replies.

Wow, reading the discussion from back then shows how much HN changed in spirit.

Even with drudge ticket-punching work, there is a vast spectrum of productivity between somebody who is capable and conscientious enough to do it correctly the first time and the person who has to rework everything multiple times because they can't do it right the first time and continue making mistakes.

That definition is too narrow. How about this one: can someone work for 1 year and (consistently) produce 10x more value for their organization than the median engineer? This seems transparently true if you've worked in any large organization.

> some niche use case that only one guy understands

But that's where all the value is! Step 1 of being an exceptional worker is not to do unexceptional work if you can at all avoid it.

> at least if we read it literally

This is where it does fall down. There isn't an X that you can reliably measure 10 of, it's all very subjective.

-1x programmers also exist, in a variety of guises. From the dev with long cv of projects he’s contributed nothing to and been sidelined from, but who manages to keep getting hired because of his cv, to the dev who crowbars in new tech where it’s not needed and causes a ton of problems.

Presumably a 1x programmer is an average programmer. The type who does his job acceptably. He doesn't create huge problems, nor solve them. He is a fine team player and team member, but not a leader. Nobody despairs when he leaves, and he will be forgotten after a few months.

Basically, the majority of programmers.

This was about minus 1x, so having a negative impact. And yes, these do exist.

Now that I think about it, the worst code seem to lead to situations where the top developers can really show their value.

I mean if the code base is quite good and is kept that way, then a normal developer can deal with it quite well. But if you get a legacy mess, then the hard-core developer can make really drastic changes that have enormous value to the maintainability, future development and bug-wise.

If a normal developer tries to tackle that mess, it's easy to get into trouble and drown in the sea of opportunities and not see the big picture and take a proper road plan to improve things.

So in a way -1 developers enable the top developers to show the gap between them and the normal devs.

This could be managers who deem themselves programmers or software architects. (Often they would use their authority to promote their ideas and effectively shut everyone else up.)

I've also met one 10x programmer in my life. A lot comes down to being able to reason fast, architect well, and write code on the fly without resorting much to:

* Documentation

* The debugger / print()

Becoming a 10x developer means that you have a set of skills that work well together in an end to end process from going to initial requirements to fully fledged architecture and/or code.

I think most people who have worked with me would consider me a 10x developer, I am capable of writing something in hours that would take others weeks (not in a half-arsed way, but taking all of the complexities of the domain into account). And I think that's what most people think a 10x'er is: that they can just do stuff quicker.

That's not what I think 10x'er is though. I think a 10x'er is someone who enables the 'average' developer to become a 2x, 3x, 4x developer. This can be through writing some core tech, a library used by all, or through mentoring and guidance. The 10x'er can then become a 100x or even 1000x developer. The 10x dev is the person who picks up the really difficult or ugly stuff and says: "Hold my beer, I'll get that done" - motivating the team and helping them to keep velocity. They lead by example and explain to the team why they've taken the approach they have.

Sometimes people use the term Rockstar to define a 10x developer. I think that's a loaded term that comes with an 'arrogance' tag. The true 10x developers shouldn't be arrogant (beyond the base level arrogance everyone of ability has), they can and should be the glue that holds a team together and helps everyone achieve more.

The way I see it:

Developer: a person that creates software.

10x development: a person that does it 10x faster.

What you describe, I'd simply call a lead developer since making other developers 2x or more is a lead developer his job.

If you're a developer and you're doing this, then it'd be one argument to get promoted to lead.

Also, if you really are a 10x dev, the one I know works in the HFT space. He's making bank. I'm not sure if that's your thing, but I've noticed they seem to appreciate 10x devs a lot more than any other company out there (and that includes FAANG).

Leading by example is not necessarily the same as being a lead developer. For example, my principal engineer is easily a 10x'er as just a pure developer, but he's also a force multiplier for everybody on the team, he's literally a machine. He leads no one, I made him principal engineer so he didn't have to.

If he was to just write code he'd be a 10x'er, with the other work he does to help his colleagues, he's many times more than that.

In the fantasy organization where means justify ends, the 10x-er's time is better spent making the 1x-er obsolete. Your lip service to the 1x-er however does not go unappreciated as keeping him on and making him feel more productive is instrumental for political advancement, which is more important for your fortunes than the technical kind.

What an incredibly cynical view of the world. Teams where people collaborate are more successful, if you think trying to make your colleagues obsolete is time "better spent", I honestly feel sorry for your co-workers. Yes, if I sat and just wrote code all day I would get a lot done. But, if I help my colleagues too, we all get lots more done.

Being a 10x developer is a factor of experience. You don’t need to debug the issue if you can already reason through to where it must be.

Yes, but some people generalize their experiences much much better and therefore can master domains and adjacent things to a degree others will never reach.

> Becoming a 10x developer means that you have a set of skills that work well together in an end to end process from going to initial requirements to fully fledged architecture and/or code.

The problem with that definition is that it tags the coveted 10x label onto the problematic cases mentioned in the article: that guy who is able to hack together POC code that he demos and passes off as the solution, but requires a whole team to fix and to rewrite and to rearchitect to get it to actually work and be production-ready and be maintainable.

Inadvertently, this means that we see the 10x label being tied to observable behaviors that are actually a productivity nightmare just because they go from zero to demo in record time.

Honestly I’ve met real 10x programmers.

Well, speaking honestly probably more like 5x, but it would take a solid 5 person team working well together to do the same things he did.

His ability to reason about the problem- get to the root of the persons actual issue, proof of concept intitial revision and iterate until it was better quality than most popular open source projects was honestly awe inspring.

His only problem was his own lack of self confidence. He used tools and always asked if they were really the best or if he was inflicting them, had good reasons for not using things that we consider industry standard (he prefers restructured text over markdown because it’s more expressive, for example) but was ultimately very sad about every project he made because he felt like he had inflicted his (good) work on the world.

So, idk. The people who think they are rockstars probably not. He definitely didn’t think he was but he was a god to me.

Disclaimer: I feel strongly about this. However, if you have a strong other opinion, then I applaud you and I'd be curious to read it. Rationally, I don't think this term matters a lot as there are better terms to describe programmers, but emotionally it's tied too much to a friend of mine and that's what makes me feel so strongly about it.

> The problem with that definition is that it tags the coveted 10x label onto the problematic cases mentioned in the article: that guy who is able to hack together POC code that he demos and passes off as the solution, but requires a whole team to fix and to rewrite and to rearchitect to get it to actually work and be production-ready and be maintainable.

As I said in another comment. I've met one 10x programmer. So I consider him to be the archetype 10x programmer since he's the only visible proof I have. We're good friends, we mostly talk about life. I sometimes see him code. Let me assure you, he is the last guy that would hack together POC code. I know this, because I'd be one of the first to do that and he gives me a lot of flack for that when I show it to him xD

I should have left out the "or" in my definition. He knows how to architect, he knows how to code, he micro-optimizes every nano, is capable of understanding how underlying ISAs are doing their thing. He knows how to reason well. At companies where he'd enter as a junior, he was immediately recognized for having better technical skills than people 10+ years their senior (note: not his social skills, those are simply normal).

They exist. And it's easy to see why 10x programmers exist. Take a bell curve of ability and you find many 10x'ers on many scales of ability.

Or would you argue that people like Einstein was not a 10x physicist, or that Euler wasn't a 10x mathematician? And there are many more that I wouldn't even know of.

10x programmer is very related to the idea of a genius, in my opinion. It's simply a subset of it.

Feel free to disagree. It's good to see other perspectives.

If their code isn't running in production with little maintenance required then they didn't actually produce the thing. Just add that requirement and it becomes clear.

So there are these moments in development, where the task boils down to ‘I just need to build this system. I can see it all in my head, and I need to turn it into code’

And for that task, factors like how well a developer knows the libraries, how fluently they can express themselves in the language, how fast they can type, how seamlessly their mental model fits with the tools available.. all of those things can make the difference between it being something one dev can crank out in a couple of days, and another dev takes a few weeks over.

And I definitely want more of the first kind of dev on my team than the second kind.

But here’s where that doesn’t match up to what we expect of lead developers and more senior devs as their careers advance, and why ‘10x’ is not necessarily an indicator of ‘ready for promotion’.

The person who can see the whole problem and turn it into code in 2 days often struggles to figure out how to scale what they do to more people. They can’t express how the code should work to someone else because they’re already three steps ahead - explaining how it needs to be done will take longer than just doing it!

So the problem is as soon as this person encounters a problem that is too big to fit all in their head and crank out in a few days, or that requires expertise they don’t yet have, so they need to collaborate… the approach they have learned to rely on stops working.

But it turns out there are 10x team leaders too - devs who can break large projects up into chunks that can be delivered separately, who make sure the team is building the right thing, who know how to make use of that individual who can just crank out a complex system in a couple of days - repeatedly and sustainably - and who make systems that are the right shape so that later features are cheaper and easier to add, rather than becoming harder and slower. In fact at this level the multipliers go way beyond 10x.

And some of those skills build on the individual 10x dev skill set but some are orthogonal to it. Some are downright opposed.

I like to say that I am a 10x developer 10% of the time. My work tends to be very "bursty" followed by long bouts of low productivity. It probably has more to do with my mental health than anything else. :/

I have the same problem. But the more I tackle it in reasonable ways, the better it gets. Here's some pointers that work for me:

- enough sleep, rather get in bed early than late as the quality of the sleep depends on natural rhythms

- engage with people that are affected by my work - this is important for motivation and to keep track of what actually matters

- operationalize everything, say out loud what to do next

- being honest when I have bad focus, talk about it

- trust my instinct and do the right thing, if I don't I quickly lose focus and motivation, if opinions don't align learn why

My personal joke about this is "I used to be a 10x developer but then I took an arrow to the knee".

Many, many years ago a colleague wrote about the experience of being bipolar and choosing to stop taking the lithium for a temporary productivity boost and subsequent crash.

> choosing to stop taking the lithium

Having been unmediated and undiagnosed for most of my life, mania is a hell of a drug. But the cost is high... really high.

Sounds like ADHD hyperfocus.

There can be that. But also some managers do not like you wander off task. So stick in your lane stick to your tasks... So eventually after years of that you just do only what is needed and skip anything else. After a few of those 'that is very nice but...' conversations you just say screw it and coast. Every once and awhile I get a burst where I can propel my group forward. But then usually I feel like I am bolder chained to them and dragging them down. More than likely I am just average and doing OK.

Was just coming here to say, this is my exact situation

I think he's beating up a straw man, here.

> The idea that someone can produce in 1 day what another competent, hard working, similarly experienced programmer can produce in 2 weeks is silly.

Over such short time scales, maybe, but the idea that someone could contribute in 1 year what another programmer wouldn't accomplish in 10 years is highly plausible, to me.

Also, sometimes it isn’t someone ‘moving fast and breaking things’ - sometimes everyone else gets paralyzed with indecision and can’t even get started.

Someone who can sit down and start coding in an ambiguous environment, period, can easily make 10x more value themselves, and allow others to go from .1x to 1-5x merely by being able to pick somewhere to start and just doing it, which allows everyone else to break their own ‘freeze’ and start engaging as now there is something concrete that breaks the locked up state they were in.

I’ve done it a few times, and seen others do it to. It’s magic going from ‘everyone rehashing things on a whiteboard for the 10th time’ to ‘everyone working on the next iteration’

The "MVP" or "strawman" is incredibly valuable; people can't tell a blank sheet of paper what they want, but if you show them a thing they'll tell you what's wrong with it.

I've noticed myself that I can happily write detailed comments on HN or StackOverflow but have no inclination for blog posts, for that kind of self-content-marketing that seems to be popular these days.

>sometimes everyone else gets paralyzed with indecision and can’t even get started.

I once worked on a project where exactly that happened, and I view it as one of my biggest failures of my career. I often wonder why I didn't just sit down one day and type int main( ... and just get on with it.

In the end the project got canned and both contractors terminated early, so I made sure that never happened again.

I like FactoryFactory devs, they are even worse than 0.1x. After some time every other team member who touches their code becomes 0.5x.

I'm not claiming to be a 10x developer or anything like that, but I have a related story to share:

Once I had to tweak a 7000 line signal generator (ie spits out wav files) written in Java. It had a mind boggling structure, and I'm sure whoever wrote it was a mad genius of some kind.

I rewrote it in less than 100 lines of Python, where each signal primitive was basically a few lines of code max, which you can functionally combine.

Since this new program was not even 100 lines long, we were able to do a LOT more with it, and in a super flexible manner.

It's one of my happiest accomplishments.

And then you encounter the 10x programmer, who is not afraid to throw the bullshit code away and write it in a way that is easy to manage for other devs. I'd say this is really a more realistic example of a 10x programmer. Most programmers are just afraid of asking the "why" question and the actually acting if there is no good answer.

> while it's inherently a super-subjective metric, I would say that 10x programmers do exist

Even worse than super-subjective, it seems like people don't agree on the definition of "10x programmer".

Is that 10x value (code * code quality) produced per unit of work-time? Per brain cycle? Are you controlling for programmer experience, or domain knowledge, as [1] pointed out? Do you actually mean 10x, or is that a stand-in for "significantly more productive" with an unknown value? Do you include collaborative skills and working with a team, or even entirely base your definition off of that?

I'm both curious as to the exact definition you have in mind, and also trying to point out the more general issue so that others arguing past each other can realize that they have two different ideas in mind.

[1] https://news.ycombinator.com/item?id=28797923

Have a 1 year project. One side is a team of 10 developers with 5 years of experience in the field. Other side is 1 developer with 5 years of experience in the field. Do you think it is unreasonable that there is any developer that could beat the team in that situation? The quality expectations for both are the same, so no shortcuts.

I don't find that unreasonable at all, most teams are very unproductive. So what I'd call it is "A 10x developer is a developer who can outperform a typical team of 10 developers".

For me, it's definitely "a stand-in for "significantly more productive" with an unknown value". Neither the input nor the output can be properly quantified, and we don't really know the full set of variables to control for, and it inevitably runs into serious questions of inequality, SES, and background that are likely to make people even more cross.

I see it as more of a compliment. Nobody should be seriously and unironically calling themselves a 10x programmer, it should be bestowed by other people.

People are often willing to sacrifice any amount of development velocity to avoid being seen to make mistakes.

IMO that attitude never makes sense - it can make sense to avoid certain mistakes at all costs (self-driving car), but it never makes sense to try to avoid blame for them (company refuses to admit their car isn't good enough).

This is a problem of culture, not individuals.

Does the culture at a company openly admit mistakes, try to learn from them, and put systemic fixes in place to address them, or do they punish people for mistakes?

If the latter, you end up with nobody admitting to mistakes and a terrible culture, it's not about the people, it's about the situation you have put them in and the incentives given to them.

No, it's also a matter of personality. One manager used to urge me to do things in half the time at 90% of quality, and stop chasing the remaining 10% at such high opportunity cost. To this day, I try, but it's hard to overcome a natural, obsessive tendency.

> it can make sense to avoid certain mistakes at all costs (self-driving car), but it never makes sense to try to avoid blame for them

Eh? As you say, it's a culture thing, and the culture of a lot of humans is that the mistake only matters if you take blame for it. Plenty of people and organisations put more effort into avoiding blame for things than the thing itself.

Blame-free culture is great, but it has to be actively maintained.

Plenty of people and organisations put more effort into avoiding blame for things than the thing itself.

Blame is pointless (particularly directed at humans), learning from mistakes is invaluable. What I'm saying is it is a terrible mistake to put more effort into avoiding blame than avoiding mistakes.

Blame-free culture is great, but it has to be actively maintained.

I agree.

> People are often willing to sacrifice any amount of development velocity to avoid being seen to make mistakes.

I'm living this at the moment and it's driving me crazy.

Maybe 0.4x and 3x programmers exist, and so when you introduce a little noise it's easy to perceive as a 10x difference at certain points in time.

> I’d rather someone give me opinions that I violently disagree with than for them to have no opinions at all.

I've been a developer for 15 years... anybody else feel like the further they get into their career, the more they want to keep their opinions to themselves?

I feel like as a junior dev I had way stronger opinions and I was a lot more vocal about them. Now, I still have opinions, but I've learned that nobody gives a damn about my opinions so I just keep them to myself.

No, I've been doing this for ~30 years now and I have really strong unpopular opinions on things. Like:

- NPM/Yarn/Gems etc is a disaster waiting to happen. Having lots of dependencies is a security nightmare and a maintenance issue. As these projects get dropped by their original maintainers we'll start seeing more supply-side attacks and more vulnerabilities going unpatched.

- Docker is a great solution for a specific problem. But I don't think it needs to be used for everything. Same for K8s. Same for microservices architecture.

- Simplicity is an important design goal for all systems. Complexity breeds bugs. Boilerplate is simple. DRY can be taken too far. Every abstraction layer leaks at least a little. "clever" code is usually not what is needed for reliable software.

I expect to get downvoted because these are unpopular opinions. But I'm used to arguing about them. It's always interesting to discuss.

Those seem like pretty unobjectionable opinions. Have an upvote!

You want a controversial opinion (on HN), how about mine:

"It's not always a good tradeoff to throw something under the bus to save developer time."

So many important things are routinely sacrificed at the altar of developer speed and comfort, app runtime performance being the big one. It's used to justify all sorts of crazy things we do in modern software to ship faster, but so hard to push back on because developers are expensive and software needs to be done yesterday.

Some things are worth giving up so that developers can move faster, but I argue not everything. Need to evaluate case by case.

Have an upvote yourself! yes, totally agree.

Like Wordpress optimises for edit functionality, at the expense of render speed. Most of the internet is now slow, optimised for Susan From Marketing because they don't know how to edit HTML.

Same for Ruby/Rails, same for Python/Django, same for Node/whatever. Web pages render slower so that whoever edits them can do that faster.

This isn't good.

Last month I ran into The Principle of Least Power, and it's grabbed me and the more I think about it, the more it expands. I think all of your statements lines up with this principle, except one.

Containers are a simpler solution than Operating Systems. Operating Systems is a tightly bound non-simplistic system that breeds complexity, just as you state npm does. Because those interfaces are tightly bound, we were not about to build on top of them. Almost immediately after we got Docker, we got K8s. Once we got Docker we got the ability to manage these things at a higher level. That's the mark of a more simple solution.

Are there things that Docker can't do? Are there things that require so much performance that Docker can't handle it? Only then should you step back to the more powerful solution.

Not to say that there aren't problems with Docker. Dockerfiles have a lot to improve. Running containers has a lot to improve.

I think. I'm still at a state where I might change my mind. It's partially why I'm posting this.

[0]: https://blog.codinghorror.com/the-principle-of-least-power/

yeah, no. Containers run on top of Operating Systems. They inherit everything that the underlying OS has. They're not simpler, they're actually more complex, but pretend to be simpler.

What you're thinking of is Unikernels. These are simpler than OS's. And yes, Unikernels are a good alternative to containers for a lot of applications we currently use containers for.

Appreciate this response. Thanks. Totally agree now that I see your perspective.

> Simplicity is an important design goal for all systems. Complexity breeds bugs

Counterpoint: some problem-spaces are inherently complex; most problem-spaces have hidden complexity. Complexity cannot be removed, it can only be moved up or down your tech stack, and obviously, you can needless add more complexity.

Docker, K8s and the like are perfectly fine when they fall below your project's complexity floor - they are good place to shift your complexity to as they are battle-tested and well-documented. If your projects complexity floor is lower than docker/k8s, then integrating them will just increase your solutions complexity needlessly.

Totally agree. Some problems are complex. However, that does not mean that the code needs to be complex. Simplicity should still be a design goal even if the problem space is complex.

I don't really agree that there is a "complexity floor". Things can always be made simpler, up to the limit of the problem space. As you say, complexity can always be added. Most projects do this. Very few stick to the minimum complexity. The complexity floor for most projects is therefore way below where that line is normally drawn.

> up to the limit of the problem space.

This is the key. One of the harder things to do in software design is to find a representation of the problem that gets closest to that complexity floor. It's often not obvious.

> I expect to get downvoted because these are unpopular opinions.

Well, not really. These opinions are very common among my peers. We sold multi-million dollar transformation projects with the stated goal of making things simpler and easier to use and understand.

Fellow 30+'er here...

You had me at

> - Simplicity is an important design goal for all systems. Complexity breeds bugs. Boilerplate is simple. DRY can be taken too far. Every abstraction layer leaks at least a little. "clever" code is usually not what is needed for reliable software.

Nuff said.

There are unpopular opinions? It’s pretty shared in the industry.

Quite brave to share these unpopular opinions like “npm is bad” on HN.

Same here. Been working professionally as a dev for 20 years and each year I care less and less about changing other people’s minds about architecture.

Just this year I was working with another team that wanted to move to an architecture that I thought was overengineered for their specific problems. They shared with me an article about some company that was using it. I explained my experiences with such a pattern and where the pitfalls were, hoping their lead dev would think about it or discuss the pros and cons. Well, it was met with silence. They went ahead with their plan anyway.

Too many times that’s happened in the work world, so at least now I don’t invest a lot of effort into changing minds. Best advice

I’ve learned these last few years is don’t tie yourself to your work. If you’re opinionated about how to do things because you love designing software, that’s great, but you’re setting yourself up to be disappointed by how things really get done.

All said and done I think you tend to learn that opinions are just that - opinions. Often a team doesn't need more opinions, they need to make decisions and come to consensus, so helping them decide on which opinion is best is more important than bringing yet another complicating-but-not-discussion-progressing vote to the table

It comes in waves. I'm in my 22nd year of developing software and I can reflect on times when I felt incredibly confident and didn't shy away from sharing my opinions, and other times when I felt like it was better to remain silent because I had things to learn from others.

There are times for both speaking and listening in our careers; as I've progressed further in my career I have felt it is often more valuable to exercise active listening.

I pick my battles pretty judiciously these days. I've learned that a huge amount of what people talk about doing ends up going nowhere anyway, so why burn the energy and look disagreeable? A lot of people with bad ideas also aren't very good at getting things done.

I think that is also because as a Junior one thinks that he should have opinions and be vocal about them and be right to further their career or to become Senior/Manager/Leader. Just like getting good grades in school you would have to put your hand up and tell that smart thing to the class/teacher.

After couple of years people realize that just having opinion and being vocal about it does not make one a Senior/Manager/Leader. There is much more to it and one learns about it on the job.

Junior dev: Let's use this amazing algo I learned in school!

Senior dev: This won't work but it's no use explaning why since they've made up their minds (as have I).

I got dinged in a 360 review by peers for not advocating for my views more.

I feel that if I present what I think we should do, and why I think it's the best alternative, I shouldn't have to get into a shouting match or even a debate. I don't take it personally if people disagree with me.

If I've got the best ideas and my managers are ignoring them for the loudest ideas, that's on them, not me.

Maybe they are pissed that they cant drag you in the failures by saying "even X had the same opinion..."

Most people mostly care about his/her own opinions, and others opinions only insofar as they reflect on himself/herself (even though of course the others are also thinking about themselves mostly), we're hopelessly self-involved creatures like that. Nonetheless there is inertia and subtle shifts that can come from sharing them, and though it may often feel like coming up against a brick wall if you look in the right way you'll see the bricks may be aligned a bit differently, and you can trace some of those differences to opinions you may have expressed.

As always the wise approach is probably somewhere in the middle, in this case choosing your battles. You don't have to voice every opinion, but you shouldn't be silent if you have a strong and well-formed opinion about something important.

Having opinions is good, but only if they’re genuinely informed opinions and everyone is willing to disagree and commit.

I’ve worked with too many people who think that their opinions are their identity and defended them as such. A difference of opinions shouldn’t become an excuse to go to war with your coworkers or argue endlessly instead of getting things done.

One of my worst work experiences was with a team of developers who were old friends from a college debate team. We could never have any productive discussions because they entered every conversation intent on winning instead of having an honest, productive conversation.

The best engineers I’ve worked with have had strong opinions but have also been happy and willing to go a different way if that’s what the team decides as a whole.

For me it's more a matter of putting those opinions into perspective. Example: I still highly value good API design but I'm not getting into debates anymore if an endpoint isn't RESTful or 100% consistent with every other endpoint. It's often just not that important at the end of the day.

I mostly express opinions about things than affect my productivity. I don't care about npm vs yarn. It doesn't affect how I work. But I care a lot about not having excessively firm lint rules because they slow down my coding.

It's the same in any area of life and is called maturity.

The longer you are in a career, the more you realize opinions are one of those things that always change (and should). Strong opinions weakly held

What a depressing reaction to adversity. An alternative reaction is to try and figure out why nobody cared, and get better at expressing yourself in ways that people will care.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact