Paul Graham has written a pro-immigration essay titled “Let the Other 95% of Great Programmers In”.
The anti-immigration people have to invent some explanation to account for all the effort technology companies have expended trying to make immigration easier. So they claim it’s because they want to drive down salaries. But if you talk to startups, you find practically every one over a certain size has gone through legal contortions to get programmers into the US, where they then paid them the same as they’d have paid an American. Why would they go to extra trouble to get programmers for the same price? The only explanation is that they’re telling the truth: there are just not enough great programmers to go around.
I asked the CEO of a startup with about 70 programmers how many more he’d hire if he could get all the great programmers he wanted. He said “We’d hire 30 tomorrow morning.”
I’m pretty sure that this is an illustration of my hedge fund manager friend’s mantra: “When the market gives you an answer you don’t like, declare market failure.” Presumably Graham’s pal could hire 30 great programmers tomorrow if he offered compensation significantly in excess of what Google, Apple, and Microsoft are paying and/or simply called up great programmers to ask “How much would I have to pay you to quit your job?” and then agreed to whatever price was quoted.
This is also an example of our philosophy around imprisoning drug dealers to end drug dealing. People are born with an innate calling to drug dealing, rather than having chosen the field due to the economic incentives presented by the market. Thus if we put all current drug dealers in prison there will be no more drug dealers. The analysis of the situation that Graham describes is similar. Just as there is no way to turn a retail clerk into a drug dealer (since he or she lacks the genetic disposition toward drug dealing) there is no way that if companies nationwide paid programmers more than the BLS’s median pay of $74,280 per year (source), additional Americans would be attracted to this field.
[Note that in my home state of Massachusetts, $74,280 pre-tax is $52,192 per year after tax (ADP Paycheck Calculator), i.e., comparable to what a person could get in annual tax-free child support following a one-night encounter with a $300,000/year earner. A person who wanted to have two children could collect more than $52,192 per year by having sex with two different Massachusetts residents, each earning $135,200 per year or more (see worksheet).]
The essay starts with a false premise and goes downhill from there. The premise (as the title suggests) is that since the US has 5% of the world’s population it also has only 5% of the world’s programming talent – that programming talent must be evenly distributed among all the nations of the world in proportion to their population. Funny then that this is not true when it comes to marathon runners or basketball players – India has around 1/6th of the people in the world and yet Indians do not win anywhere close to 1/6 of all Olympic medals.
Also in the idea that “great” programmers are born and not made and that in all 300 million people of the US there are not an additional few thousand people who could be trained to be great programmers, so we must import them from all the nations of the world in proportion to the population of each country.
Also, it’s really a pity that we don’t have a system for transmitting files and communicating electronically so that a programmer in some distant country would not have to uproot himself and his family and move to San Jose but could work in his impoverished native land and contribute to growing the middle class there. Someone should invent such a system. I have an idea for a system whereby packets of digital data could be routed to remote locations, each identified by a unique digital address, where they could be reassembled into files but I don’t know whether such a system is really feasible. If such a system could be created (and I know this is a big if) it might even be possible to set up audio and video links where you could meet face to face with programmers in distant lands without having them emigrate to the US. If (and I am really speculating here) such a system existed, it might be possible to “source-out” programming work to other countries. Maybe in the distant future something like this will exist someday.
Sigh.
To quote from Graham’s article:
“What the anti-immigration people don’t understand is that there is a huge variation in ability between competent programmers and exceptional ones, and while you can train people to be competent, you can’t train them to be exceptional. Exceptional programmers have an aptitude for and interest in programming that is not merely the product of training.”
Basically the argument you have with Graham is whether or not there is such a thing as a “exceptional” programmer vs. a “competent” or “average” one, and whether such variation can be explained by education. Graham argued that it cannot, and you seem to argue that it can.
Graham on one hand, is a person who is an accomplished computer programmer, and is surrounded by accomplished programmers and claims that there is such a thing as “exceptional” programmers, and that such aptitude does not spring merely from more education.
You on the other hand, a Harvard Lawyer (of unknown programming provenance) who seems to claim (without any sort of evidence) that any such distribution of excellence can be explained merely by education (inferred from the said lawyer’s drug dealer/store clerk education analogy and the fact that the lawyer quoted from the BLS MEDIAN performers salary, not any salary for exceptional programmers). which point of view would the rational person believe?
The one from the informed source, perhaps?
Jason, let us accept your contention that excellent programmers are born, not made.
Are all of these prodigies actually programmers? Or does their programming ability lie dormant in some cases, as they have pursued other paths, like banking, medicine, law, the arts, nonprofits – paths on which they may earn more money, social prestige or both. They are like stem cells that have chosen to mature into a different kind of cell.
The flaw with your analysis is that presupposes that everyone who could be a programming prodigy already is one. A rise in prodigy programming wages and prestige would turn more of these latent programming prodigies into actual, active programming prodigies.
Similarly, paying Olympic athletes a lot more might divert the nation’s top athletes from say, the NFL to Olympic competition, increasing our medal count. (Whether we *should* do that is a different matter.)
@Jason programmer with 8 years and a above average salary chiming in. Graham is just plain wrong.
Why?
While there are exceptional programmers in the world they are a minority (miniscule). Good programmers are trained with investment of time and energy. How do I know, I have been through it, and so have 50+ programmers I have worked with in the past decade. Just like doctors are not magically born with medical knowledge and are trained so are good programmers.
In the end its all about the bottom line. A programmer (and I was that programmer 6 years ago) from Pakistain or India are exploited by undercutting a US worker since they do not have the power of negotiation. They are paid 75% of the salary of a US employee (see h1b rules if you are not sure why!)
But if it is true, what about other professions? Doctors, truck drivers, lawyers there are plenty of other professions whose 95% too are born out of US should they have special visas too?
I think what we are missing is an organization that protects the rights of programmers, I do not hear anyone wanting to import lawyers….
My humble opinion: you can’t train people to be passionate about something. Being an excellent developer means being best in the world at a constantly and rapidly evolving set of technologies. You need to be all in, all the time.
What is the size of the population that earns less than $74,280 in your state? why are they not “attracted to this field” as your argument claims they should be?
Jason: you picked a poor choice of subject and author for an appeal to authority. Paul is neither an accomplished programmer not surrounded by them. He has written no software of note ever, and he surrounds himself with delusional startup weasels, not programmers.
The author of the blog you are commenting on is this guy: http://en.wikipedia.org/wiki/Philip_Greenspun
So if you want to just shut your brain off and not think about the issue and let “the informed source” decide for you, then you are betting on the wrong horse.
carlsberg:
To answer your question, some people who are not attracted to the field but could do well in it may be working as math professors or nuclear physicists. If Paul Graham’s CEO friends were willing to pay $500k/year, those people might leave their current fields and go into programming.
Jason,
Your ad hominem argument is baseless. Graham and Greenspun have comparable bona fides.
To quote Wikipedia:
“Greenspun … received an S.B. in Mathematics from MIT in 1982. After working for Hewlett Packard Research Labs in Palo Alto and Symbolics, he became a founder of ICAD, Inc. Greenspun returned to MIT to study electrical engineering and computer science, eventually receiving a Ph.D.”
“Greenspun released the software behind photo.net as a free open-source toolkit called the ArsDigita Community System, built on top of the Oracle relational database management system. He wrote several textbooks on developing Internet applications, including Philip and Alex’s Guide to Web Publishing, SQL for Web Nerds, and Software Engineering for Internet Applications, the textbook for an MIT course. Greenspun started a company to sell support and service contracts for the toolkit, which remained free, and grew ArsDigita to about $20 million in revenue before taking a venture capital investment.”
Regarding Paul Graham’s ideas, it is evident to me (16 years in the software industry) that the executives just don’t want to pay the elite talent what they deserve under any rational accounting. Grahm writes “it’s easy to imagine cases where a great programmer might invent things worth 100x or even 1000x an average programmer’s salary” but he never talks about paying a great programmer 100x or 1000x an average programmer’s salary.
The big picture here is not that your startup is competing for talent against Google and Facebook, it’s that the entire software industry is competing with Wall St, Med School, Business School… If you are a top student you will realize that the big money is not in software engineering.
Carlsberg: What is the size of the Massachusetts population that earns less than $74,280 per year? http://www.bls.gov/oes/current/oes_ma.htm says that the median hourly wage is $21.07 in Massachusetts. Let’s assume that the programmer works 3000 hours per year to earn his or her annual salary (most active coders seem to work more than 40 hours/week). A worker at the median wage in MA working 3000 hours per year would get paid $63,210. Thus we can say that approximately 60 percent of Massachusetts workers earn less than the nationwide BLS median programmer salary.
Why wouldn’t all of the folks who get paid less than a programmer want to go into programming, which does pay more than average? For many it would require an expensive investment in education (where the expense is college tuition plus foregone wages during the time spent in school). For others it would be that they prefer the jobs that they do have, e.g., because of the people they meet while working or the stuff that they get to do while working. At our flight school we have had programmers quit to become flight instructors and then professional pilots, for example, because they were tired of sitting at a desk all day, they enjoyed the people they met in the aviation world, and they enjoyed flying, despite the lower pay and higher risk compared to software engineering.
I remember an ugly approach that I made once to the big airport in Toronto. As I was beating myself up for making so many power adjustments the captain said to me “Nobody was born knowing how to fly a 53,000 lb. jet.” (see https://philip.greenspun.com/blog/2013/07/11/asiana-214-training-with-passengers-in-the-back/ ) I do think that (wise) pilot’s perspective would be useful to Paul Graham. Nobody was born knowing how to write C code. Somehow they were induced into reading Kernighan and Ritchie. A higher salary is a powerful inducement and in fact that is more or less what Graham is proposing. But instead of proposing a higher salary for people who are already resident in the U.S. he is proposing to offer higher-than-current salaries to people who are living in other countries. So at some level Graham is not disagreeing with standard economics. He is just saying that the incentives should be held out first to non-Americans.
[On this point, I got a private email today from a Uruguayan who works remotely for a U.S. company: “I would hope the immigration reform would allow me to move to the US and work directly from there… Currently I make U$S 20/hr and it’s a very nice figure for living here, but I know a job there pays twice as much…”]
@jason: I would use the word “Paul Graham” and “accomplished computer programmer” sparingly in the same sentence. His claim to fame is that he wrote a shopping platform in the Lisp programming language, then sold it to Yahoo which enabled him to become a VC. As someone who used to program Lisp professionally (AutoCAD), Lisp is a horrible choice for commercial programming. Eventually Lisp code devolves into an unmaintainable mess because of the odd features (eg: dynamic self-modifying code, untyped unstructured data, etc.) that I’ve read Paul praising. Yahoo eventually threw out all the code that Paul wrote, then rewrote it at great expense in a more mainstream programming language.
I’ve worked with many H1Bs in several Fortune 500 companies. Despite what these companies say to the press and the govt, at the ground level it’s all about getting cheap labor to fill seats. Based on the quality of their work, I can’t say I’ve ever encounter one of these “great programmers” working as a H1B, but I have seen entire departments staffed with dozens of citizens fired then replaced by much cheaper “temporary” contractors from low-cost centers (aka LCCs, corporate speak for India) brought in on multi-year contracts.
http://www.idlewords.com/2005/04/dabblers_and_blowhards.htm
philg : So we agree that 60% of the people to whom these opportunities are available do not train themselves (you do not need to have an expensive education to be a great developer — look at the Linux kernel core devs), move out of the profession because its hard/taxing, or make more money than the jobs offer.
Not sure I agree with the piloting analogy. The software industry changes way too rapidly, you cannot be a recreational programmer and be classified as a excellent. Again, this is about _drive_. Maybe the Uruguayan who emailed you is really, really, good and is a game changer that helps company A beat company B in the marketplace, and would bring an entrepreneurial drive that very few have.
(I should share that I’m an immigrant tech worker, in another G10 economy though. Nothing beats the Silicon Valley startup engine, and guess what — one day I’m going to be there.)
Sorry read that as “and the other 40% make more money than the jobs offer”.
Graham’s post is downright autistic. He seems to think that the employer’s advantage should be the sole determinant of immigration policy.
Foreigners imported into any country have large effects on the entire community they are in, serving to reduce trust and social cohesion. We not only endure the social effects of immigrants, but of all their descendants, possibly forever. By the principle of reversion to the mean, the descendants are unlikely to be excellent programmers.
@Carlsberg
One migrant tech worker to another. lets face it the technology does not really change that rapidly and you do not need to be a savant to know programming. Case in point, how many times have design patterns really changed in the last 5 years, heck they are identical across c#, java and any other oops oriented language. Sure you have your nodejs, your go but its not the toughest job in the world.
The Uruguayan in this instance should and would love to work for 60,000 a year but you can bet that a programmer making 80,000 would be out of a job. Now is the American developer any worse probably not, he most likely is more experienced (after all he was hired for it) he just costs more.
End of the day its all about saving a buck folks.
Oh and Carl we will welcome you with open arms whenever you make it, I was!
“practically every one over a certain size has gone through legal contortions to get programmers into the US, where they then paid them the same as they’d have paid an American. ”
Right, but not the same as an American “rock star” programmer, which is what these top talent hires essentially are.
So basically the industry ARE trying to drive the salaries down, from the top, and Graham is telling half-truths to tilt the board in his favor, just as we should expect since he is a VC.
The operating assumptions in this entire discussion seems to be that American programmers are paid appropriately, or are a little underpaid, for what they do.
That’s like saying European and American manufacturing workers in the 1970s were paid appropriately for what they did. Maybe they were from a moral perspective, but people demanded cheaper goods than they could produce.
The analogy may not be perfect for the US, but in Europe, we had extensive ‘H1-B’ programmes for manufacturing workers. Like Paul Graham’s proposal, they were supposed to keep our industry competitive; mostly they didn’t, because in the long term it makes no sense to uproot a Turkish employee, move him to Germany, pay him 300 DM a month, when you can move the factory to Turkey, and pay him 30 DM a month.
For some reason, programmers expect their ‘education’ (99% of which is actually available online, for free) to exempt them from the same market forces that destroyed the western manufacturing base.
Michiel,
It’s one thing when market forces result in the inevitable, which is the relocation of work to places that have a comparative advantage, but it’s another for the government to actively collaborate with industry to the disadvantage of its citizens. Most governments, even those that are generally in favor of free trade, do not allow free movement of labor. It’s reasonable to ask whether bringing in foreign workers (and their grandmothers and cousins, etc.) results in an overall improvement to the long term well being of a society (as opposed to short term improvement to corporate bottom lines obtained at the expense of their native born workforce). As the European experience with the guest workers shows, cheap labor and the ability to get currywurst is a dubious trade off for long run major disruptions to the social cohesion of your society. There may be legitimate needs to bring in true “superstar” performers (e.g. Linus Torvalds) but the truth is that most employers are just looking to save money and they would like to bring in armies of worker bees who will work cheap and not just a handful of true superstars.
Wow, wow, wow.
You all miss the central argument in the article, Philip included:
“We have the potential to ensure that the US remains a technology superpower just by letting in a few thousand great programmers a year. What a colossal mistake it would be to let that opportunity slip.”
It’s advantageous to live in a country that has unfair economic advantage over other countries, like US right now. Now, once we over the moral dilemma of collectively exploiting others and decided not to move to Africa, next logical step is to preserve this advantage and enjoy high standard of living. As such, same as we do with scientists, we want to syphon great programmers from all over the world, basically to corner global market in this precious resource. I.e. deny other competing economies great programmers, ensure availability for your own.
Now, it’s actually unclear of it will drive down programmers salaries: programming market is driven by global demand, which is still rising, if all world programmers reside in US, they would still be in demand. Right now we just limit growth of US software companies, and allow growth of competing software companies by not relocating talented programmers.
SK: You correctly restated part of Paul Graham’s argument, but not the part that Greenspun responded to.
Greenspun is taking issue with Graham’s claim that the technology industry isn’t trying to use immigration to drive down American wages. That’s something Graham said in his article, Greenspun quoted it, and as Greenspun has said above: it’s not true.
> Now, it’s actually unclear of it will drive down programmers salaries: programming market is driven by global demand,
The market for programmers isn’t worldwide, it’s not even nationwide. A programmer working in Palo Alto will make more than a programmer working in Austin. One may argue that the market could be or should be worldwide, but the fact is that it is not.
> As such, same as we do with scientists, we want to syphon great programmers from all over the world
American employers recruit scientists from all over the world in order to drive down the wages of scientists in America:
http://users.nber.org/~peat/PapersFolder/Papers/SG/NSF.html
Izzie, I don’t know where you get your information about Europe from, but most people will happily trade these supposed ‘major disruptions to the social cohesion of our society’ for a single Kebab sandwich.
But back to the topic: Graham’s proposal is phrased like a friendly ultimatum, just like the European manufacturers’ was in the 1970s: “We want to stay here, but if you don’t give us what we want, we’ll move abroad.”
Many Americans seem to believe their country’s immigration policies are liberal and generous. They are not. Just as an example, the Netherlands’ H1B programme is about 1/4 the size of that of the US, while its population is less than a twentieth. The wealth disparity between Eastern and Western Europe is as great as that between the US and Mexico, yet labour movement is completely unrestricted.
Tariffs and regulations will not work in any way to contain an industry that needs no physical infrastructure save the internet, but slashing labour costs might help a little while, to keep a very labour-intensive industry from doing the logical thing and relocate to where labour is cheap.
Don’t buy Graham’s argument that the effect of this policy would be for the US to remain a technology superpower; an industry dominated by 24 year-old JavaScript programmers with Bachelor degrees does not, in any way, represent high technology. This is good, old-fashioned corporatist labour policy. But it might work for a little while.
In this discussion, we are neglecting that there are visas like the EB-1 and EB-2 which give people with extraordinary talent the opportunity to apply for a greencard. This is outside the H1-B cap and is truly aimed that attracting the best to the US.
http://en.wikipedia.org/wiki/EB-1_visa
We have plenty of programmers (especially modestly skilled ones). If we want to siphon off the super-productive superstars from other countries, put a wage floor of $300k/yr on h1b visas.