Part of this is the fault of government - in almost all statistics the $200K systems architect at Google is rolled into the same column as the $45K coder pounding out VB6 apps for the company intranet.
This article is even worse than most, since it lacks any real data on programmers, and instead uses data for all STEM fields. This is a little like measuring the average height of people in Texas using data from all over the world. It's just lazy.
The more accurate title is: "top level developers in extreme shortage, mid level developers less so, USA swimming with underemployed low level developers"
And this reflects the reality on the ground. You have companies like Google, Facebook, Microsoft, et al, who are desperately trying to bring in foreign labor at the $150-200K levels (and beyond, I'm starting to see comp packages towards $300K), showering them with lavish perks that the rest of the tech industry hasn't even heard of, much less the rest of America.
Then you have the flip side, with companies like Infosys and WiPro bringing in cheapo labor at $50K a pop and renting them out at exorbitant rates, using the terms of the H-1B to hamper and exploit their employees knowing they have little recourse. You also have domestic low-skill IT workers who, despite exploding demand, have not seen their wages rise significantly.
Anyone who talks about tech immigration in the US who doesn't acknowledge these two sides is either woefully uninformed or has a vested agenda. The reality is that we need to expand immigration in the former category and hit the brakes on the latter, but this also doesn't fit nicely into the "immigration is purely a means for exploitation" narrative that some people are trying to spin.
Schools are very good at producing crappy programmers, and not very good at producing top-notch programmers. This is not a surprise, nor is it unique to America. The "look at all of these unemployed STEM graduates" argument is useless without also looking at whether or not said STEM graduates actually qualify for the jobs they seek. Simply shoving a shiny, rolled-up piece of paper at someone doesn't automatically make them competent.
The reality that we need to attract more top talent, and the reality that companies are lobbying for more cheap labor are not mutually exclusive.
That said, the shortage of top developers could be attributed to any number of things, including:
* There is a limited number of top performers in all fields. This is natural, and cranking out more college grads won't change anything.
* Top performing students are choosing other fields, perhaps ones that pay better (finance, medical, consulting, basically any field with a higher glass ceiling)
* Or they choose jobs with better perceived job security. You don't constantly hear about outsourcing, indentured employment, and 10-years-to-irrelevance in most fields.
* The students lack the pre-requisite education to excel in CS (this probably isn't a huge problem among the top performers, but you can't avoid the fact that the elementary education system generally sucks and many prospects probably fall in the cracks).
* Companies (and by extension, our law makers) are taking a short-term outlook by prioritizing the immigration of lower-tier and middle-tier coders?
Vs the article
"134,780 H1-B visas that were approved in 2012."
A good analogy to the debate is we're talking about a shortage of, and reasonable desire to import any legendary epic level multi-platinum award winning rock star, but in actual practice we're importing anyone who's ever touched a guitar in order to keep domestic wages low by pushing unemployment higher than it should be. They are two separate issues.
134,780 H-1B visas were approved in 2012, what is the distribution of salaries? The fact that the US let 134,780 people into the country last year says absolutely nothing about how many of these people are cheap, barely-qualified foreign labor, and how many are the multi-platinum award winning rock stars. It's quite obvious to even a lay observer that both categories of hiring are occurring in large amounts in this pool, but almost no one is willing to address the balance.
Painting the entire H-1B pool with a single brush is, to be blunt, extremely ignorant of how it actually is in reality.
And seeing the state of the industry, a simple mean would be an inaccurate picture also, given the extremely lopsided distribution (i.e., lots of people making $200K, lots of people making $60K). What would be far more useful is an actual, comprehensive look at the distribution of H-1B visas. The total count is meaningless except in the populist "raaaah our jerbs!" way.
Ah I think I see our minor disagreement. My belief is the vast majority of the eighth of a million per year are NOT elite rock star caliber based on some crude estimates about graduation rates and so forth. I don't believe we have the educational capacity, world wide, to graduate 250 million computer science grads every 20 years outside the USA such that the elite top 1% from each graduating class year eventually H1Bs over into the USA. Even if we really did have 250 million outside the USA CS grads every 20 years, why would a perfect 100% of the 1% come here?
The percentage of elite hires almost by definition must be so low in that eighth of a million, so as to be a rounding error approx zero. That doesn't mean Mr Anecdote doesn't exist or they're evil or imply any political ridiculousness, it just means that elite level coders are such a tiny fraction of the population that shoving them into the debate sideways is intentionally derailing the real debate about the vast majority of people involved.
Beyond discussing the present and past, if we imploded the H1B system from an eighth of a million to perhaps two thousand, almost certainly the two thousand who actually survive the process would almost certainly be the elite 1% supercoders, so even pretty far out "solutions" would have no effect on supercoders, making it even less relevant to debate supercoders.
Another interesting way to put it, is most countries allow anyone in if they post enough money. I don't know what the USA charges, but $1M would be more than any other country I've ever heard of. So for a mere $2B, if H1B were utterly eliminated, you could still let all the supercoders in under the wealthy people rule at minimal cost. After all, if you're willing to pay $350K/yr for a supercoder, posting $1M for a wealthy person immigration is only about three years total cost.
Talking about the tiny population of supercoders is simply a distraction.
Probably none. The actual rock star level people can get O visas, which don't have a quota.
Funnily enough, immigration into the US has traditionally been to fill lower-class working roles. The slave trade was the first and most obvious example of this, but we also have Chinese, Irish and southern European immigrants in the 19th century. If immigration is what made our country an economic powerhouse (and at the very least, the slave trade was one of the primary contributors to our capitalist development), then we should be encouraging all immigrants, not just those who can be immediately identified as "strategically important."
You're going to get the amateur historians all wound up with talk like that.
A better explanation was the Whitney cotton gin enabled mechanized processing about a century before the Rust mechanized harvester was invented, so you had explosive growth in a couple strategies for providing human cotton pickers ranging from ownership pre civil war to indentured servitude / peasanthood after the civil war. The gin wasn't enough to kick industrialization into high gear, although I'm sure it helped a tiny irrelevant bit. By the time they (the people) were made obsolete by the Rust Harvester, and shipped up north to the factories, we already had the worlds biggest factories to ship them to... so much for that argument. The south has never amounted to anything from an industrial standpoint until very recently and even that's a bit of a stretch.
Or if you were alluding to the earlier triangle of unhappiness (whatever it was called, triangle of something miserable anyway) then we had pretty much maxed out molasses/rum/human transportation well before the industrial revolution. The limiter was like how much land can you cultivate in the Caribbean and how many slaves can you ship there, they didn't have industrialization in the new england rum distillation industry as a limiter to growth at that time so there was no pressure for improvement. Distillation is not much of a motivator for industrialization anyway. Every couple years I alternate between brewing/distilling doesn't get enough credit for industrialization and thinking they already get too much credit.
A pretty good argument overall, other than getting the relevance of slaves in capitalism all wrong.
Many "good" programmers I know would very likely jump ship at their current job for 600k/year. You've just got to find a better middle ground.
17 is just a year away from being able to get married or join the army. Those seem like a bit bigger deal, especially since you can switch majors before graduating (for example my first roommate, who switched from CS to... something to do with business I think).
The incentives here are incredibly obvious to spot. Facebook/Google can either pay a star developer 600k/yr for the duration of his stay, or bring a foreign indentured developer and train him for 60k/yr. The choice here may not be obvious, but the reality is that both of these strategies are employed.
Most top tech companies have executives which are paid 10 to 100x the base developer salary, because they manage a hive of indentured foreign workers who can do the menial dev work. These executives essentially become "Architects". This is the best case scenario, in the worst case, the architects are also trained or outsourced.
The economics of this situation are very simple to understand, and explain why most companies employ the dual strategy to significantly reduce their costs. Green card holders have the option of negotiating for more compensation as they gain experience, but foreign workers without status do not have this option. As such, they are a far better bargain over the long run, assuming both are of equal ability as fresh new grads.
If you are a GOOD developer, the opportunities for you are endless. And the bar for being a "good" developer is NOT very high relative to the range of abilities found among all software developers. Even horribly incompetent developers do not seem to have much trouble finding work (they do have trouble finding work at this company... usually).
I think the reason why roughly one third of "computer and information science" graduates aren't working in their field is because they themselves chose not to work in this field. The work isn't for everyone. I know quite a few people who majored in CS, were able to grind through it to get their degree, but can't stand the thought of doing this for a career. Their motivations for going into the field out of high school vary, and many are misguided, but how can you really hold a 17 year old to making a 40 year life decision?
We need more good developers. If we were allowed to hire more international developers, I can tell you we would not hire one crap developer "because they are cheap." A crap developer does not just do less quality work than a good developer, they do NEGATIVE work that costs the company more than the salary difference between them and a good developer.
From a self concerned strategic standpoint, the U.S. should use the opportunity of immigration reform to steal good developers from other countries. We are a country of immigrants, and immigration will always be important to us. Let's use it to capture the best and brightest from the rest of the world. That is how we stay on top.