Skeptical. Remote work increases the supply of available technologists across the board. That means more competition at every level - 10x or otherwise. If your whole team is remote, what exactly is the difference between someone in California and Chile?
Another thing not mentioned here is that employee productivity is a variable that is almost always invisible before the hire happens. Leonardo DiCaprio can command a premium because he is identified with the work he does. This is not the case for tech workers who's work is typically invisible. Suppose you are a 10x DevOps sorcerer. None of your potential employers will know that in advance, so they are all only willing to pay 1x salaries.
FAANG companies pay higher across the board, not because they want to, but in the hope that by overpaying 90% of their employees (relative to market) they will be able to attract top 10% candidates whose skills are hidden until after the hire happens. If they could price discriminate in advance they would.
Don’t necessarily agree with the article, but I’m a little curious about this. If you don’t have anything to show that you are a 10x DevOps sorcerer, are you actually one?
The thought is that it would show through your past experience, no?
Just because someone meets the "X years doing Y at Z" criteria doesn't mean that they are great hires. I've met PhDs in computer science that could barely write python.
Resumes or similar self-marketing says a lot less than we would like. If there were good signals here, hiring would be a solved problem.
FAANG companies don't treat their extant employees anywhere close to well enough for me to believe they're trying to attract top 10% candidates by overpaying 90% of their employees.
IMHO building teams that work together well is far more indicative of success than individual high performers, and as a result the economic impacts will not happen to individual "10X" performers but rather teams, organizations, and companies that can harness the network effects (i.e. what we see right now, where companies like Facebook, Google, Amazon, Apple, etc. are more dominant than individual people).
I don't think it's such a mystery or revelation as the author seems to make it out to be.
When a country/people are young and poor, like we were in the 1950s, everything is about physical goods to make our lives better. That's where everyone is spending their hard eared money, incurring their labor. Manual labor, to make goods and services -- and those services that don't transmit very far.
Now, 60 years later, everyone has huge material wealth built up -- the legacy of decades of labor and productivity. So first of all there's just a lot more money sloshing around to be spent. Money is spent easier now -- do you remember how tightfisted your parents were about money? Because it was scarce! Now it's easy -- even they are more casual about it.
And then, what is there to spend on? You don't need another dishwasher. So you spend on leisure activities, luxuries, intellectual/social goods. Things that don't require huge amounts of manual labor. Things that can be recorded and replayed, and everyone is willing to pay a couple dollars for.
The people who produce things for that market are of course gonna be the ones who luck out.
But the pendulum swings back and forth, and I'm sure it will again for us some time. Go to another country, and the pendulum is on the opposite side.
The consulting work I do delivers things in a day that would usually take clients' staff 10-20 days to do, and there are many things I do in a minute that would take most people in my field an hour.
I can do this because security skills are a very long tailed distribution. Does that make me 10x to 60x?
> The 10x programmer is a myth, it's just that corporate collaborative software development practices routinely produce teams of >10 that get less done than a single competent developer working independently
My theory is this is one of the main benefits of hiring a consultant. They can cut through bureaucracy.
Yep, some of the most added value I have seen in consulting was essentially taking recommendations that department A was unsuccessfully requesting for a long time already and getting department B to actually implement them, i.e. an organizational shortcut enabling something that simply wasn't possible given their current structure.
Pretty much every large company has a lot of good things that they aren't doing because of their dysfunctional internal bureaucracy that causes them to disregard good proposals (and sometimes intentionally sabotage them for reasons of internal politics) until they are re-packaged by some expensive consultants for top management blessing.
Agreed; my dad interviewed to be a VP at Lockheed in the 1960s and eventually turned down the job saying, "All your problems are political! None are technical!"
At my previous job, I used to sit in meetings thinking all the problems were political/communication, and those are the only interesting ones.
Over the years, I didn't come to the conclusion "I want to be insulated from non-technical problems" (which is an easy thing to achieve) but rather "I want to be involved in solutions that are technically trivial, but solve organizational issues". Because those are the ones where the most leverage and meaning exists.
It is necessary but not sufficient to be a legitimate 10x technical subject matter expert. Most organizations lack the expertise but they also lack effective leadership and organization to apply the expertise even if they have it. Part of the job of being a highly paid “10x engineer” is to fix the organization so that the organization can fix the technical problem. The former will consume 90% of your time.
Some 10x engineers have legitimately exceptional technical abilities that stand on their own, but the critical skill for many (most?) is being really good at smoothly fixing broken engineering processes and organizations from first principles. A handful are good at both.
The main benefit of hiring a consultant is that management can blame them for uncomfortable changes.
Anecdote time: I requested a CI runner from the systems/infrastructure team, and they ghosted me for two years. I just stood up a CI runner myself using my own resources, about a week after they ghosted. Recently they hired a third party firm from India to make a CI/CD pipeline, never told me, and claimed credit for "the first CI/CD pipeline at the company".
A career in infosec tends to develop an extremely broad set of skills (with commensurate lack of depth in many of them) that can turn you into a generalist superhero when shit hits the fan. Back in my heyday I would get called in to help fix outages as much as I would for security work. Feltgoodman.
Ten years later and I'm now adrift in some weird role between an engineer and program director with zero accountability. I just go from Zoom meeting to Zoom meeting knocking down bullshit and connecting the right people in a big company. I've essentially become a giant tube of organizational lube. Feelsbadman.
Indeed, the breadth piece is huge. The depth, so long as you have some extra self awareness that you literally only know the table of contents level of most things, and can do some composition of them while knowing where to find people with depth in them, you can solve a lot of problems. If you can tell whether you need new science to solve a problem, or it's a function of work by composing a bunch of solved problems, that's going to elevate your abilities fast as well. I wonder if the right metaphor is security is like the eng.sci of tech.
Reality is in security is we've spent a career navigating, negotiating, and implementing the real power relationships that the technology facilitates. These are performance, economic, legal, business, organizational, environmental, institutional, etc.
The article contrasts "10X" performers with "mediocre" performers, but this is exaggerating the required differences.
The difference between an Olympic medal and fourth place is often just hundreths of a second. Similarly small differences in luck can mean the difference between being classed as "10X" and being classed as "mediocre".
This is not the first time I'm hearing the difference between a an olympic medalist and a 4th place is just a hundredths of a second. The difference between a 10x and 1x is orders of magnitude. That will be the difference between an Olympic medalist and a college athlete. Here's a video that demonstrates such gap. https://www.youtube.com/watch?v=i93vF0WOX6w
There are careers that are winner take all. Like artists, movie stars, professional athletes. If you are in these kinds of careers you are either struggling or doing fantastically well.
Then there are the professional white collar careers. That pay above average salaries and employee millions. accounting, engineers, law.
Can software development really switch to winner take all? I have my doubts.
The 10x class here are not 10x programmers - they have 10x or more the reach because they can be a celebrity/leader on a platform that reaches vastly more people. But the other side of that is the very few people will be able to command much attention, 1 in 100k or something. They will suck up tons of money and the network affect of being able to reach thousands of people at once will suck ever more money to them. As people lose their jobs from increasing automation, the somehow lacking in clear conclusion article says more income will go to these kind of incredible-reach people.
If I only have 10 coworkers, how much does it matter if someone is capable of reaching 10 times as many people as I can? That might be 8 vs 10, instead of 8 versus 80.
The TikToker has more or less direct relationship with the viewers and customers (advertisers). Developer who does not own the customer relationship is interchangeable and will not be paid on a star level.
I think people read too much into it, the reason some are paid more than others is simply their appeal, and that's measured subjectively. It's easy to see some tiktok account has more followers and is worth more. It's hard to see the worth of an SDE, that's why 10x itself is called a myth, 10x of what? 10x the lines of code? 10x the influence to upper management? 10x at cutting corners? 10x the mentoring of others? 10x your revenue stream? 10x your customer base? The linked study says some people were 10 times faster at solving a maze, is that who you want as your CTO?
Is “Baumol’s cost disease” just a fancy term for the effects of supply and demand in the labor market? I do not see how it is anything more than a trivial result of supply and demand and why it merits a special name.
Why deny credit? Plenty of economists before Baumol (and not a few after) failed to appreciate this phenomenon. As the name suggests, it's a phenomenon about cost moderating the effects of supply and demand, and it's clearly not a trivial result of analysis involving only those two. It's worth understanding in its own right, and AFAIK is typically taught for that reason in Econ 102.
> Baumol's cost disease (or the Baumol effect) is the rise of salaries in jobs that have experienced no or low increase of labor productivity, in response to rising salaries in other jobs that have experienced higher labor productivity growth.
To me, this seems like the same as
1) the supply of people willing to provide labor x at price y is reduced because the same people have an opportunity to provide labor z at a price > y.
2) supply of people willing to provide labor x for price y is reduced, therefore buyers have to pay a price > y for the same amount of labor x
I do not understand what makes this a surprising phenomenon or which “cost” is moderating the effects of supply and demand. People have better opportunities, therefore supply of labor is decreased, therefore assuming if demand does not go down proportionally, then prices will rise.
For some jobs, demand will not keep up with the rising price and the number of jobs will greatly decrease or cease to exist, such as horse buggy drivers.
You're missing a step. Workers didn't have an opportunity to provide labor Z by random chance. They had that opportunity for a specific reason: greater efficiency gains (i.e. reduced cost per unit of production) in Z than in X. In other words, something about another profession affects wages in X. This is different than intrinsic changes in X, requiring analysis beyond supply and demand for X.
Somebody had to look at this phenomenon from outside that narrow micro-economic perspective before it became "obvious" (only in retrospect) to others. That happened to be Baumol and Bowen. Why not use one of their names as a convenient handle for a concept they studied and brought to others' attention? We give "fancy" names to algorithms and even whole systems in computing too, even when their details can be similarly derived from first principles. Do you think we should shun those in favor of purely descriptive names as well?
I have no qualms with using names for it, I am curious why this specific application of supply and demand, which seems obvious to me, was designated a name.
As an illustration, if milk was just discovered to be useful to make cheese, and the demand for milk rises proportionally more than the supply of milk, then the price of milk would rise. As far as I know, there is no specific name for this “phenomenon”, but there is if you substitute labor for milk?
Edit: I sort of see the significance if the claim is that increases in productivity in other sectors will necessarily lead to decreases in supply of labor which will cause prices to rise overall, but is that true?
I can envision a scenario where increases in productivity are putting more people out of work quicker than it is causing people to be in demand, causing wages to stagnate or fall (in real terms)? It might even be happening now.
That is literally what Baumol and Bowen wrote about. I suggest you look to their papers for the evidence. Even those who posit that Bowen's Rule accounts for a greater share of cost increases in education (trivial to find many papers on this) accept that Baumol's Cost Disease exists. It might not be necessary or happen all the time - nobody but you claimed such - but it's definitely real in some industries or professions.
Is it possible that increases in productivity in other professions can reduce demand for labor in other professions (automation), increasing the supply of workers, which then suppresses wages in other professions? Would that not run counter to Baumol’s cost disease?
I appreciate your time in engaging in this discourse. I’m not sure if I will have time to read their papers, but I was interested if there was a simple explanation for why I am wrong that it didn’t pass my initial smell test.
It's intended to capture the idea that wages rise in an industry despite per-employee output not rising, which is against the "marginal wage equals marginal production" basis of classical Marshallian economics.
Ironically, given the examples in the article, Baumol's example of cost disease was musicians.
> against the "marginal wage equals marginal production" basis of classical Marshallian economics.
I am not aware of Marshallian economics, but I was under the impression that prices (wages) moved along supply and demand curves, and do not have anything to do with production.
There are different ways of figuring out what a wage will end up being, and looking at how they interact (which is what Baumol's Cost Disease does) can be tricky but enlightening.
Supply/demand is one such way. Another is related to elasticity. If a given job can produce $1.00 of value but the next person qualified for that job only expects $.90, someone should be expected to create that job and capture the surplus. And this should continue until the value produced by the next such job is right at the value produced by that job. So you often expect that the wage of a job should be the value produced by that job, at least at the margin.
Baumol's Cost Disease shows how when one job's value increases, the wage in jobs that require similar skills can see a growth in wages beyond the degree to which it's value increases - that is, the first job following the model described above causes another job to stop following that model and instead be lifted into a supply/demand model.
So concepts like efficiency and elasticity don't matter? The supply and demand curves for different goods and services varyfor specific reasons. Those reasons are worth studying, and sometimes naming.
> If “imperfect substitution” is so high for software developers, why is their income distribution far less skewed than that of entertainers? Because of geography.
> While music and TV stars can be broadcast from and to anywhere, software engineers can only work in one place.
The author understands that SWEs relocate right? And that while remote work is "modern" many companies are completely internationally located.
The thesis here is based on equating the mythology of the 10x engineer with top athletes and entertainers without considering the demand models and performance measurement.
In sports, billions of people compete from childhood on a physical competition with exactly the same rules where performance is completely objective and immediately visible. People only want to watch the very best, so as long as you have global distribution of live feeds, then of course all the money is going to very tiny set of top performers.
With actors/musicians, it's not 100% objective, but it's still directly based on audience appeal, and you can pretty quickly determine who has popular appeal and draw a through line from various business models to what the superstar can economically be paid.
However in the vast majority of businesses, it's a team effort there is no direct connection between individual contributors and customer choice. The closest role for this would be sales, but even in that case they're selling the product and the brand, not themselves. Furthermore, effective contribution usually requires a fair amount of ramp-up and domain knowledge. As a hiring manager it's incredibly difficult to know who will be a 10x-er before hiring—even if you have reliable inside information from other companies, you don't have a guarantee it will pan out in your environment. Certainly there are some superstar engineers with specific well-known expertise, and they do get outsize offers, but there are very few such roles that require "the top global expert in X niche" and while engineers can get 7 figure offers for these, there's not enough of a through-line to revenue for it to rise to the level of superstars in other disciplines which are more or less a direct function of the size of global attention and discretionary spending they command.
Another thing not mentioned here is that employee productivity is a variable that is almost always invisible before the hire happens. Leonardo DiCaprio can command a premium because he is identified with the work he does. This is not the case for tech workers who's work is typically invisible. Suppose you are a 10x DevOps sorcerer. None of your potential employers will know that in advance, so they are all only willing to pay 1x salaries.
FAANG companies pay higher across the board, not because they want to, but in the hope that by overpaying 90% of their employees (relative to market) they will be able to attract top 10% candidates whose skills are hidden until after the hire happens. If they could price discriminate in advance they would.