Why is there a shortage of talent in IT sectors and the like?

There have been some good posts on this lately, for instance asking why the wage simply doesn’t clear the market, why don’t firms train more workers, and so on (my apologies as I have lost track of those posts, so no links).  The excellent Isaac Sorkin emails me with a link to this paper, Superstars and Mediocrities: Market Failures in the Discovery of Talent (pdf), by Marko Terviö, here is the abstract:

A basic problem facing most labor markets is that workers can neither commit to long-term wage contracts nor can they self fi…nance the costs of production. I study the effects of these imperfections when talent is industry-specifi…c, it can only be revealed on the job, and once learned becomes public information. I show that fi…rms bid excessively for the pool of incumbent workers at the expense of trying out new talent. The workforce is then plagued with an unfavorable selection of individuals: there are too many mediocre workers, whose talent is not high enough to justify them crowding out novice workers with lower expected talent but with more upside potential. The result is an inefficiently low level of output coupled with higher wages for known high talents. This problem is most severe where information about talent is initially very imprecise and the complementary costs of production are high. I argue that high incomes in professions such as entertainment, management, and entrepreneurship, may be explained by the nature of the talent revelation process, rather than by an underlying scarcity of talent.

This result relates also to J.C.’s query about talent sorting, the signaling model of education, CEO pay, and many other results under recent discussion.  If it matters to you, this paper was published in the Review of Economic Studies.  I’m not sure that a theorist would consider this a “theory paper” but to me it is, and it is one of the most interesting theory papers I have seen in years.


There isn't a shortage of IT people. There is a shortage of that very specific guy you need with that very specific skill in that very specific thing that you have a short deadline for.

Which brings us back to the issue of why firms are so allergic to providing training to people who are in the ballpark with their skills, but may not hit a grand slam. This didn't used to be such an issue; in the 90s you could jump from skill to skill and pick up what you needed on the job. I went from doing VBA coding to VB6.0 without much trouble. Why can't someone nowadays go from VB6.9 to VB.net in similar fashion.
And don't get me started on those employers that list twenty "required skills" so detailed that only someone who has done that exact job for them could qualfy.

And don’t get me started on those employers that list twenty “required skills” so detailed that only someone who has done that exact job for them could qualfy.

Those are often pro forma job postings that will be used to justify hiring an indentured low-salary H-1B.

Exactly right.

Nailed it on the head.

HR departments don't know how to screen for "someone very smart who can learn new things quickly."

I've never met someone in HR who was interested in talent, only filling roles. A list of specific technologies is common because that's what HR wants.

That's what HR understands. Most HR people aren't particularly bright and certainly not knowledgeable about the positions they're hiring for. They just want a checklist.

HR recruiting roles is one of the biggest drags on a company. Except for low end, hourly wage workers I've so rarely (if ever) seen HR do a good job of recruitment.

They do: IQ tests. Unfortunately, they aren't allowed to give them anymore.

IQ tests are not predictive of employment success. And they do have bias issues, whether you are willing to admit that or not.

However testing in general is certainly not forbidden or absent in the hiring process. I have had some sort of test for most of the jobs for which I have been seriously considered. Often these were specific skills tests; sometimes they were general aptitude tests, rather like the SAT, though not as lengthy. And once I even had to take an utterly obnoxious "personality test".
But testing is very limited in determining how people will do in long haul-- let alone how ethical, compatible etc. they are.
Moreover testing (like background checking) is usually done at the interview level with just those candidates who seem like a good fit. There are practical reasons for that of course.
What the "skills checklists" do is winnow the stack of resumes down to a relative few who can be given a closer look. And while we all love to bash HR, it isn't just HR that thinks this way. More than a few hiring managers do not enjoy that role, and they too would just as soon have only a tiny number of possible hires to interview.

The problem is training is expensive, and many software development shops already have to do lots of training on the in-house code and environments even if the new person has the right basic skillsets. Even if you hire a nearly perfect "fit" in terms of language and general skillsets, it'll take at least a few weeks before a new hire can understand the code well enough to make meaningful development contributions. If they're "off" by a lot in terms of skillset, bootstrapping them to the level where they can both understand the underlying tech and the product itself could take many months. This may be OK in a large corporation supporting an in-house system, but startups can't afford the cost in time and money.

Also, if you *are* willing to hire someone without the base skillset, you tend to either an obvious rockstar or you hire a smart young kid. You rarely are willing to hire a middling mid-career type seeking to make a career change (even if they'll take low pay) since they may not succeed in making the transition, and firing a dud is far more expensive than leaving the position unfilled.

How do you know what the new set of required skills will be? You don't, and by the time you do, its too late.
Plus you have massive invested capital in other fields, and employers see that very negatively. I have a friend who can't get hired as an accountant (despite just getting an accounting degree) for example, because she worked in management for too log. Employers see her resume and refuse to hire her, as she is very qualified for a management or HR job, but would be a newbie accountant.

VB6 to VB.NET is enough of a change that you might as well train someone to use a real language like C# or Java. Yes, I went there.

And to answer your first question, as someone involved in the hiring process, I'm _always_ up for training people who are in the ballpark with their skills.
I hired a designer/HTML guy to do design but also to train him up in SharePoint and workflows. I pushed for hiring a contractor who just finished a CS degree
instead of someone completely new because I knew his work habits and attitude.

VB.NET is, for all intents and purposes, identical to C# but with slightly different syntax. They compile down to usually identical CLR bytecode, and someone fluent in C# or VB.NET can easily jump from one to another.

Indeed, I'd argue it's a better, more expressive language than Java.

Yeah, I went there. Language flamewar for the win!

Sorry, but if it can't be complied into something that runs native on a Linux machine, it's not a "real" language! Good day sir!


I'm currently on a VB6 project here in Denver. The SAP replacement is already 1 year behind schedule, and won't be ready for 18-24 more months. So the 14 year old VB6 app has to be kept limping along (and because it was supposed to be replaced last year, there is no business case to migrate it to anything). Meanwhile, the CTO and CIO alienated all the full-time programmers so they quit, and that leaves contractors like me to carry on. Normally, I reject VB6 contract assignments from recruiters, and I don't even mention it on my resume anymore, but I like this recruiter and they were desperate (more desperate than usual).

>Why can’t someone nowadays go from VB6 to VB.net in similar fashion...

All the VB6 programmers I know have done so. And won't go back to VB6 except to port it to .NET (with the exception of this job). This means that all experienced VB6 programmers are in their late 40s and older (I am old enough to do catch-up contributions). You can't buy VB6 except maybe on eBay, and Microsoft quit making it available to download years ago. I don't think that any VB6 book is still in print any more. Even COBOL is in better shape for long-term future survival than VB6, especially with ABAP (SAP's scripting language) being a derivative of COBOL.

Finally, the stigma of VB6/VBA has endured enough that most .NET developers only want to work with C#. I think that VB.NET and C# are pretty much the same, but some things, like LINQ are much easier to use in C#.

>in the 90s you could jump from skill to skill and pick up what you needed on the job...

I know one guy who does this still - learns what folks are looking for during the interview and then tries to learn the stuff before starting the job. This ticks off the folks who hire him and he's gotten a bad enough rep here in Denver he had to move away (to DC where he can still get away with it). Software development isn't really a large field - no matter where you go, you will meet folks you have worked with before (or at a minimum work with people who have personal friends who worked with you).

This was my initial thought as well. We've seen an incredible expansion of computer-related knowledge in the last 40 years. Hiring "a computer guy" is like hiring "an engineer." What do you want, a bridge built or a radio designed? So, when you say "IT people" to day you probably imply more specificity than you know.

I agree with this. We are long past the day of needing a "computer programmer" for any kind of advanced work. How many types of engineers does an engineering college graduate and how many types of computer programmers does the equivalent college graduate? Generally, programmers aren't differentiated very well and programs vary widely in rigor.

We should also note that computer fields invent new specialties at a high rate. The invention and then demand for smart-phone apps developers was a recent example.

I was writing smart device apps in 1999 and 2000 for the RIM pagers (before they even had phone capabilities, before they were even called "Blackberry").

There were a lot of smart people in that space trying to work with the RIM or hack external network connectivity onto a Palm or iPaq 3600.

The iPaq 3600 seemed to have a cult following.

Right, but speaking of "IT" and employment, perhaps some of your 1990s co-workers took a wrong turn and weren't resume-ready for those phone jobs. Tyler is a smart guy, but he may not get how many jobs go empty because applicants have "the wrong IT."

May not? He certainly is clueless about the actual realities of finding a job. Essentially all economists are.

That makes sense to me. The biggest issue I see in recruiting talent is a severe lack of people that understand complex systems. It isn't the skills per se but the application of those skills to ambiguous problems.
To make matters worse, few managers really understand complex systems and there are no training programs to help people acquire these skills, you either understand or you don't.

No training programs for complex systems? I'd say that's exactly what my engineering degree was. Granted, I specifically took the super-computers class for that exact lesson. The specific skills I learned have been and probably will continue to be useless for the business tasks I do.

Funnily, in the 6 years or so that I've worked with "supercomputers" there's very rarely that I've met anyone that has had a formal class on super-computing.

For almost everything I do, development wise. I have never had a class for.
You just learn as you go along. Taking a class would simply be too slow.

Oh wow what insight - companies hiring in technology are all chasing the same small pool of veteran workers at the expensive of bringing in new people - I believe I've posted numerous comments about this very phenomenon. Oh but I don't have a PhD in Economics so I guess it doesn't matter.

You should go be a comments section star at some place like USAtoday.com You could benefit from the color coding and people will argue with you all day. Have a blast.

CBBB, with all due respects to your insights, the paper Tyler is referencing here is an actual peer-reviewed and published paper. Would you care to submit all or even some of your comments to peer review?

But my point is that if THIS is what passes for "one of the most interesting theory papers I've seen in years" in Economics - then that's a pretty sad statement of the discipline. Look at this paper he's doing all this modelling, doing all these manipulations for what? To come up with a very trivial and obvious conclusion?

Arm chair economics is not helpful. The theory may be the end result of this work, but the models are where the real interest lies. There are many people who can name this problem, but none of us can proscribe fixes because
a. We have no logical proof to back up our claims. This is the purpose of the peer review system.
b. We have no audience for our ideas, even if we have a real fix.

I think many of us are drawn to this forum because maybe, just maybe, we can influence Tyler or Alex's work. Perhaps, CBBB, you are part of the reason that TC finds this theory paper interesting.

Please do not feed the troll.

I think it's pretty depressing that this kind of thing is considering insightful. It's a legitimate point.

There is a lot of conventional wisdom that is just plain wrong. Occasionally we need to empirically test what we think we "know." We sometimes have to model reality to understand its nuances.

We might suspect that the decision to commit a crime involves some form of cost-benefit analysis, but the factors that contribute most to that analysis - to an individual or in the aggregate - are not a clear-cut as we think. The consequence of the research is that we gain insight into public and private measures that work and don't work.

The vast majority of published research is esoteric nonsense. It's somewhat refreshing to see research that speaks about things that really matter.

Would the article be more insightful if the conclusion was contrary to conventional wisdom? One of the biggest problems in academic research is that negative results are published much less often than affirmative ones. The irrelevance of the research and the obviousness of the conclusion bothers me more when it is funded with taxpayer dollars.

Lack of loyalty has its costs. For firms, they must overpay for mediocre, experienced talent. For workers, they must cover the entire cost of their training. My own sense is that firms are in the best position to address the lack of loyalty, but, for the most part, management incentives are simply too short-term for them to care. If you get paid based on what happens this year, there is almost no reason to consider spending a lot on loyalty initiatives and training workers.

Loyalty is a cost. The constant evolution of IT means that anything that remains stable is by definition not a complex system, or is not considered worth the cost by business. Thus, the only long-term IT skills a company needs are maintenance of stable or non-critical systems (as defined by business, we in IT often disagree). So loyalty means small pay increases.

In my experience outsourcing has destroyed loyalty in many firms. This is firm specific. Those that have an outsourcing strategy are looking for rationales to move work overseas and they do not want to train newbies in USA at all.

>The constant evolution of IT means that anything that remains stable is by definition not a complex system, or is not considered worth the cost by business.

I think it depends upon what you're talking about. A large number of skill sets in IT remain valuable over long periods not because they are complex, but because they are flexible enough that they can solve a wide variety of problems. The issue with IT from the business side is that a service or product can quickly move from a cutting edge "low-hanging fruit" profit maker to something more marginal. However, alot of the changes that management seems to hear and understand are coming, not from IT People, but from software sales companies that want to sell you the latest VB.net or Oracle solution even if you don't need anything more complicated than an open source LAMP system.

KLO, that's a well-reasoned post. I'd add to it the statement from Chris, above, that "few managers really understand complex systems" - which is hardly a situation that is likely to inspire loyalty or respect for the people who work with those complex systems.

An interesting contrast would be an accounting firm. Accounting (particularly for corporations or taxes) is enormously complex, and these firms support young talent in acquiring the CPA credential (and turn over many who can't). The differences might be explained by the fact that accounting is a much more unified field than is "IT", and the managers of accounting firms are accountants themselves, who tend to understand & respect the value of continually investing in their workforce (and earn high enough margins to afford it).

It also helps that public companies are required to meet very strict reporting requirements and hire outside auditors. An IT project can always be deferred, but not an annual report.

Technology evolves with such speed that younger IT workers are generally worth more than older IT workers.

The question I don't get is why older IT workers don't keep up out of fear. Yes, a semi-conductor engineer is going to be miserably forcing himself to learn xcode or java to build mobile apps, but it isn't like it will take him 4 years, and he brings a strong knowledge of power trade-offs to the math behind the apps he builds which gives him a leg up in the market - apps can't just be cool, they do need to worry about power consumption.

And again, a semiconductor engineer is going to have a far easier time with codeacademy than my musician niece...

We need a system that messages to those in declining skill-sets that they need to get their ass in gear.

That is a ridiculous assumption - and I'm a "younger IT worker".

Did you even read the summary of the blog post? It reaches the opposite conclusion - no one wants to spend money on young and unproven talent.

Young in this context means 25-35. Everyone wants students to get training somewhere else then hire them.

But young guys gets hired in IT. The old guys have trouble finding jobs. His model must be horribly flawed if he predicts the opposite.


Oops, there I go again, forgetting to read!

You might be faced with a biological problem. Fluid intelligence peaks at about 25 years of age and then continuously declines. Re-learning is a hard business.

This is the best suggestion I have heard so far.
It makes the most sense too.
Also programming isn't like most other disciplines. Every time you do something you are doing something new. Its like functional art... with math. Every job requires lots of fluid intelligence.

I call bullshit. Learning is only difficult if people stop learning. The advantage is always to people who choose to learn their entire life.

I think biology has some to do with it, but much of it has to do with priorities. I've recently made a big move overseas and had some time on my hands (since I'm not allowed to work here yet), so I learned iPhone programming. It isn't that different from most any other programming system I've had to learn, but the fact is it still took me 2 weeks straight just to get the basics down and another four or five weeks to program my first app for sale.

I can't think of very many companies that are willing to set aside large chunks of time for anyone to learn a new system from scratch, and older workers have things like jobs, mortgages, and family to make such a huge career move very difficult.

I didn't start seriously programming until I was over 30, and didn't write what I consider to be professional level code until a couple years of hacking around. Now I'm a CTO.

Why do I have a feeling I've never heard of your company?

Who cares? I've created tens of millions in shareholder value. Investors and customers have heard of us. That's all that matters to me.

In a lot of cases he has money already. I know quite a few people retiring from IT in their 40s and 50s.

As someone who as gotten his ass in gear over several iterations, let me assure you that it takes more than this alone. Depending which analysis you read, the IT market both has a problem with trusted young talent and it also distrusts older talent. I increasingly find that my talents are geared more toward companies that can't afford downtime rather than companies trying to break into new markets with new techniques. As much as I enjoy new market development I realize that my experience doesn't match their immediate needs. Maybe not all senior developers are good at reading the signals, but rather than describing it as a problem of training I would say that they need to know where their skills are needed better.

Tech firms are gaming immigration rules. In order to justify more H1B visas, they have to prove they cannot fill positions with qualified Americans. So they have a perpetual pipeline of unfilled positions.

I think the visa rules are inane. First, it's a pointless exercise to try to "prove" that there isn't a qualified American available. I don't see any practical way of settling this on a case to case basis; only thing you get is reams of forms.

In any case, a paper-pushing immigration bureaucrat is probably the least qualified to judge the merits of the case.

I agree, but what is the alternative?

We had open borders in this nation when a shortage of labor was our biggest concern. In many places, those new immigrants were exploited by natural born citizens and early immigrants. They became fodder in gang wars and political battles.

Now, as in any other advanced economy with a full to partial welfare apparatus, our duty is to sift immigrants through a fine filter as long as we have a surplus of domestic labor.

When the whole world is your recruiting ground, the top 1% of foreign candidates will be more attractive than mediocre citizens. That would be optimal if and only if our mediocre citizens were willing and able to emigrate to places where they are the top 1%. Other nations have immigration controls, so we must too.

Every position occupied by a foreigner deprives a US citizen of that opportunity UNLESS no American meets the minimum qualifications. A nation must make some sacrifices to serve its citizens. Call it an inefficient birthright privilege instead of a pure meritocracy. The resultant is always a question of what the medium- and low-skilled citizens will do to earn a living. Why are 50% of graduate programs filled with foreigners when American college graduates are getting rejection letters?

The more signalling plays a role in hiring, the more I believe that Americans can be trained to adequate proficiency in a lot of jobs, both high and low tech. The "skill" and "intellect" of foreign candidates is also sometimes suspect. They often come from the privileged aristocracy of their home countries. We should wonder why we import so many foreign low-skilled workers when we have a large quantity of low-skilled Americans who are trapped by the gills in our social safety net.

Our incentives have distorted the choices for young Americans. Remove these incentives, and the ship will begin to right itself. You can't convince me that there are many smart young Americans who want to be computer programmers and the like.

>>> but what is the alternative? <<<

Here's one idea: Have firms bid for visas out of the total quota each year; or maybe have top-level sector specific quotas at most. Get rid of the charade of the firms "justifying" each hire to a jobsworth bureaucrat who gets to play God about who America needs and who it doesn't.

> You can’t convince me that there are many smart young Americans who want to be computer programmers and the like.

Did you mean to say "are not many"? If you meant what you said (this would seem a bizarre statement IMO), then can you perhaps
add a bit more about what "the like" is, in your mind?

I hire computer programmers all the time, and for the freshest graduate hire we make (which most of them are ... hmm, this is not very consistent with the applicability of the paper to IT) no one is going to make less than six figures - right off the bat. Exactly how much "incentive" are you thinking about to get these potential top 1% American's to show up?

"I hire computer programmers all the time, and for the freshest graduate hire we make (which most of them are … hmm, this is not very consistent with the applicability of the paper to IT) no one is going to make less than six figures – right off the bat."

Maybe you made a mistake here but you're saying no one you hire (including fresh graduates) make less then 6-figures right off the bat? You're obviously full of it - I know people who work at Google, RIM, Cisco, etc. as software engineers and they don't make 6-figures.

In certain geographic areas I can see that six-figure minimum applying.

That sounds reasonable, for a consulting firm.

CBBB -- the big companies don't necessarily pay the most, especially at the lower levels, that's one of the more common IT myths among new entrants (in fact their perceived stability and clout enables them to pay less).

I concede my situation isn't typical (there are geography and industry specific variables) but yes, I very regularly hire (am a senior person involved in the hiring process), and for the completely fresh grads we regularly hire compensation would not ever remotely approach $100k (i.e. would not be anywhere low enough to approach it).

The comment I am responding to talks about the "top 1%" of candidates. This is who we are after. It would be (even more) presumptuous of me to say we can identify those reliably (relevant to the initial thread! Bonus points!). But yes, even after the identifiability problems, a candidate who is likely a top-somewhere-%-ile software engineer, even if right out of college and with no industry experience, isn't going to be offered less by us. If they stand a good enough chance to be reasonable calibre, why would we pay them so low - just to exploit them?

I also have passing but less intimate experience with silicon valley startups. A proven software engineer (yes I am shifting the rules a bit here) taking under $100k, without what (she believes to be) a very high equity upside - this isn't exactly ridiculous, but I'd bet it to be a distinct minority.

And to follow up (what I think is a very thought provoking comment by Willits) - if I find a great candidate from elsewhere in the world (seems surely top 1% or better), I really want to hire this person. It will probably be H1-B. Do I care about "prevailing wage" provisions and such - no, it would never occur to me to think about this; if they are a steal at $150k (and top 1%, that would be a steal) I'm not going to risk the hire. Of them many many (CBBB skepticism invited) H1-B's I've hired, if the US Government were to have imposed a say 20K/year tax on the hire (hint, hint, USG, revenue opportunity here!) I would suspect at the end of the day no hiring decision or salary offer would be impacted in the slightest.

A deeper question: if we can poach the world's top 1% in area X, is it good for us? For those we poach from? Is such efficiency in the geographic allocation of "talent" a good thing at the end of the day? I can step back and doubt it, even if - if there is a problem - I've been part of the problem. I'm more conflicted today than yesterday.

You are correct. I meant to say "aren't"

Google doesn't pay as well for entry level because getting a job there right out of college is a meal ticket for the future. A close friend of mine works in HR there and has done tons of hiring over the last 5 years.

People with master's and PhDs do get paid very well, however.

"I hire computer programmers all the time, and for the freshest graduate hire we make (which most of them are … hmm, this is not very consistent with the applicability of the paper to IT)"

I think the paper is claiming that it is difficult to enter the IT field when you have complimentary skills and high potential, but not the exact career experience or skillset needed for this particular job in this particular company. In my case, I have a BS in mechanical engineering, had a brief career as a civil engineer cut short by the implosion of land development. Meanwhile I've taught myself a few programming languages and enough web development skills to offer those services freelance, all as a hobby on the side.

When the construction market shrank for the foreseeable future, I couldn't get an interview in IT without a litany of credentials, hands-on experience with enterprise software, and all these things that the capable, educated and self-starting technical professional can't get without having had *that* IT job before. So at 30 (I'm certainly not old by most standards) I took my unemployed self through another BS, this time in IT, at a cost of $18,000 in student loans. But it got me an internship, which is the foot in the door I need. As a new IT grad I shouldn't have a problem getting interviews. It goes without saying that I'ie learned way more in my intenship than in all those classrooms, and my company is happy with my contribution, by all accounts. But why did I have to be a member of the class of indebted university students to get a shot at a new field with on the job training? For all they know, a company could be missing out on some unique benefits of a technical professional who enters the field a different angle.

New grads in technical fields are always valuable because they are cheap and they are easily "programmed" in the ways of the hiring company. Career changers have to overcome this disadvantage, and for me it comes at the cost of paying for my own proper training (more precisely, credentialing).

(In full disclosure I work for an IT consulting company)
Willitts, But where are the qualified Americans? Certain skills, yes, you can find them. But my company has years of history posting for, say, a J2EE position and rarely Americans applied. Furthermore, if we do a search in one of the major job boards, like Dice or Careerbuilder, it is rare to come across citizens for some of these roles (so that eliminates our company as being the source of the problem for lack of applications). Instead we end up rummaging through page after page of “H1B visa holders” or even OPT visa holders claiming to have a H1B visa. So where are the Americans hiding out?
I hardly think the preference is to avoid American workers in order to fuel more immigration. We also have records on everyone we submit to clients and who is selected. These records will show that, on average, for every x number of candidates submitted, the very large majority are likely to be visa/GC holders. The remaining submittals are Americans. Yet, in most cases the Americans submitted were ultimately selected. This leads me to believe that at the end of the day, the hiring managers prefer the Americans (or hopefully they are simply selecting the best qualified for the job).
On the topic of training - Companies are providing on the job training whether they know it or not. A good percentage of applicants completely fake their resumes (not embellish- but fake it outright). Half my recruiters’ job is to play detective to see if the resume is for real. So many tech workers now go to a two week course on a technology and get their friends to be references for them. I’m sorry to say, but often times these are foreign labor workers who do this. Since I rarely come across Americans with the highly technical skills, I cannot speak if they are doing this en mass too. I read so many complaints on American companies creating abuses to favor larger immigration pools. If this is happening, they are not the only ones abusing the system. Some candidates and the company that holds their visa are also abusing the system by promoting them to fabricate their work history. I’m not against the H1B program as whole, but these abuses are ridiculous, especially when companies say they can’t find the business case to provide training, yet that is often exactly what they are doing at the end of the day.

"But where are the qualified Americans? "

They are either suing people, doing boob jobs or inventing some kind of synthetic financial instrument.

Any american raised kid that's smart enough to figure out where the money is will not be doing IT, much better, getting into Law, Plastic Surgery or Investment Banking. The guys actually getting into IT are usually geeks with low social skills, or guys that are not bright enough.

Way off topic but is there an open HTML tag in the above post?

It looks like it. Lets see if this fixes it...

A lot of comment systems match <i> and being ended by a <i/>, even though they shouldn't.

Bender probably tried putting </i> in his, but the comment system stripped it out because it wasn't matched up with a starter tag. (Imagine if he put in <BODY>.)

I can emphasize or bold or strengthen or italicize but the browser is unlikely to close the above tag. <!---->nor is this.

It's really a bug in the commenting software.

PS: I hope this doesn't come out all ugly because I messed up the encoding.

Need moar </i>'s

Yeah, that's the really crazy thing -- American tech workers actually get a pretty substantial premium just for being American. The language/culture barrier is huge. If you can smooth out your personality and have decent work habits you're golden.

And yes, the number of fake tech workers is truly awesome.

If there are really all these empty J2EE positions out there, then how come I have applied to hundreds, and no company will contact me? It's hard for us job seekers to believe that companies really have these shortages.

I am a .NET developer (in Denver), and I don't use Dice, Monster or Careerbuilder. I found those 3 to be less than useful looking for real jobs, and so I haven't used those in 8 years. Nowadays, I used LinkedIn to post my resume as it results in far less spam, and so far the jobs listed on LinkedIn are not fake jobs where recruiters are trying to collect resumes. My anecdotal experience with J2EE is that Denver mostly lacks them, and the jobs in that area that do exist are typically contracts where wages are lower than for equivalent .NET jobs - because one is competing on price against H1B visa holders. I'd rather compete on value.

Maybe your city is a Java city, and there are a lot more Java jobs there. That isn't the case here in Denver.

Yes so we should give up trying right Rahul?

No, we should look for one of those rockstar programmers everybody is talking about. /fn !i \!i )off @@term NO CARRIER

No, tech firms legitimately have difficulty finding the right workers for their task. H1Bs are the default because American IT workers demand more than companies are willing to pay for someone who isn't necessarily a great fit. For many reasons tech firms want American workers, but they don't want an American ZMP. An H1B ZMP is easily removed and probably cost the company much less. So it's the decision between the right American worker or an H1B. We need a systemic change for this to be different. And it has nothing to do with H1B visas.

"H1Bs are the default because American IT workers demand more than companies are willing to pay for someone who isn’t necessarily a great fit."

Prove it. This is inane. I think there are tons of young engineers out there who do not get on the career ladder and would be willing to take lower salaries.

The real issue is that companies don't want the burden of an American workforce.

If two guys (one American and the other not) are equally skilled and demand the same wages what incentive does a firm have to not hire the American? Explain the "burden" if it isn't wages nor skills.

I think there are tons of young engineers out there who do not get on the career ladder and would be willing to take lower salaries.

We didn't find them last time we were looking for them. Most have very high "self-esteem" and even higher expectations. The ones willing to take low salaries are pretty useless.

" would be willing to take lower salaries. "

Define low salaries. Will they relocate to middle America for $40K per year? If not then they are not willing to accept low salaries.

How is it easier to get rid of a native ZMP worker? If the HR is worried about lawsuits, it seems that H1-B would just create more legal hurdles.

I'm going to assume you meant harder in your first sentence, since that's what I said and it jives with your second sentence.

I've never heard of legal problems with H1B termination. All I could dig up was that the contract must be paid in full unless contract was breached (as per any contract employee). The only H1B specific piece I could find is paying for transportation back home, which shouldn't be a burden for anyone hiring H1Bs.

What are the barriers to firing native workers that don't exist for H1B workers?

Are H1B workers, say, prohibited from suing for unlawful termination? That would seem to be the major cost with canning a native worker. Otherwise give them 2 weeks severance in exchange for agreeing not to sue and be done with it.

In terms of programmers for a few specific languages some have theorized that many recent (in the last couple of years) acquisitions of small start ups by big firms has worked more as a talent acquisition than anything, particularly when the the purchased business is shut down. The 'talent' has shown that it can perform the complex tasks and the loyalty is ensured (for a short time) by partial payment in unvested stock.

Can you provide an example? I am very curious now.

Been there before myself. Compamy gets eaten for the talent, which is offered an extra year worth of salary for staying two years.

What the acquirers forget is that it is not just talent, but culture. If your culture is no good, people will still jump ship the minute it becomes economically advantageous. Put your typical Agile crowd to work under a waterfall organization, and you will lose most of their productivity even if they stayed.

The examples are endless. I would bet money that a large majority of all acquisitions by Twitter, Facebook or Google are talent acquisitions - I know of at least two cases personally, and it's usually really easy to tell from the press releases.


A large percent of this list looks like Facebook buying up possible future competitors more than anything else.

That's a small goal. Facebook hasn't been worried about competition for quite some time. The workers are exactly what they want. If facebook wants a new feature, it can either get people together and determine how to implement that feature, or they can hire people that have already done it. The latter group could go from concept to production much quicker, saving Facebook money and potentially retaining good talent.

Beyond raw talent, you're also getting a team that has worked together for a while creating something new with limited resources. That's very valuable.

Facebook has done a few of these. FriendFeed and Gowalla come to mind.

Seems like an explanation that works even better for the Finance sector...

That's exactly right. The skills of the incumbents are even more highly valued when they come with a book of business. In fact, you don't have to have much skill at all as long as you bring your clients. Unless, of course, the desired "skill" is the retention or accumulation of clients.

>>> why don’t firms train more workers<<<

The shortage is for positions where you just cannot take an average unemployed guy and get him up to speed in a year's worth of coaching. The skills needed are hard to master quickly; and the people who could master them are already employed elsewhere at better wages.

e.g. In a typical engineering major less than 10% of the class is good in programming; and the others will probably never ever be good; no matter how much you worked on them.

That's true, but I wonder how much of that is from a lack of diligent effort in our education system and misaligned incentives for potentially gifted American students.

Suppose you have high school and university teachers who are not the most-qualified for their jobs. The poor high school teachers are protected by their union, and the poor professors are protected by tenure. They are permitted to be lazy and arrogant, focusing their attention only on the most gifted American students and allowing the others to fall by the wayside.

Then we have a government which subsidizes fine arts, liberal arts, and sciences which are not in high demand. Students are drawn into degrees that have no marketable worth. It's a monumental waste of resources.

I wanted to be, and could have been a gifted programmer, but the education system gave me poor teachers and professors. It wasn't that they weren't smart - it was that they were apathetic, arrogant, abrasive, and unaccountable for their results.

I switched from computer science to economics as a major. The prospects for an economics major looked poor, so I went to law school. But many other students got those same signals, and being a lawyer wasn't as high paying and intellectually stimulating as I thought it would be. So I changed careers into financial services where I managed to survive drastic cuts because I was marginally better than the people who got pink slips.

I do well and like my work, but I would rather have been a programmer. I would like to program in my spare time, if I had any. And I'm so far behind the curve that I don't think I would be successful even if I won the lottery and didn't have to focus on anything else.

I agree with almost all of that. All I'm saying is that the IT talent shortage doesn't have quick-fix solutions. "Why don't the firms train more workers?" is a silly question.

Your example is arm-chair nonsense. Many of the chemical engineers were better coders than my computer engineering brethren. They had to do a lot more of it than we did. I wouldn't hire them for general purpose coding, but they all could easily pick it up.

I didn't even mention an across major comparison. All I'm saying is a good programmer (irrespective of major) is a rare commodity and it's not all about training.

You said a lot more than that. You even gave a percentage without any facts.

Good programmers are everywhere. Good programmers who know technology A, design pattern B, and company C's data and organization are rare. And teaching everything at the same time is extremely costly. If that's what you meant, then you are correct. But that is far from what you said.

This is the sort of insight for those who've never looked at the Sunday employment ads for professionals in the obscure paper, The New York Times. You know, the large boilerplate ads that communicate, "There are only three people in the world that can manage this supply chain and they all work for our competitors so here's the offer we'll make to lure you away."

This is the kind of paper that makes people think that Economics is not "the dismal science" but "the science of BS."

Part of the problem is that in IT mistakes are very costly and very very very hard to spot by anyone, much less high/mid management who typically can't even read the work product of the programmers much less really grok it.

So this is really a management problem? I agree. Solution?

How do we turn good programmers into good managers and leaders? Are the skill sets the same?

How much do we sacrifice when we take an outstanding programmer and turn him into a sheepherder?

Once you stop programming, you stop being a good programmer. After 5-10 years this quickly makes you a terrible manager of programmers.

Maybe if you had some sort of system where developers worked together in small packs, with a lead that had maybe a third of his time devoted to management?

This is appealing, but I think that undervalues how much management and product oversight is needed for a given unit of development.

It should be closer to 50% for the lead, with each other developer doing 20% management work (or work that's a substitute for management). The issue is, developers tend to hate management and management-y things. It's not typically in our skill set--oftentimes the personal factors that drive someone to being a good developer cut against being a good manager--and even if it is, managing is really shitty compared to actually building something.

You mean agile programming?

Your questions are too general. Any answer has caveats.

It does not help when the mismanagers read fad articles and books with titles like "Does I.T. Matter?" with the argument that I.T. does not matter. These things just continue to justify to the pointy haired bosses that they don't have to know what is going on at all.

I'm reminded of an article: "On Being Sane in Insane Places" (Rosenhan, 1973), where researchers got themselves admitted to psychiatric wards. The "groupness" of the folks running the wards was enough that they ignored the sane researchers because they had pegged them as "insane". The researchers were even taking notes on what was going on, and were not hiding the note-taking because they were considered crazy already by the folks running the institutions. Unlike the folks running the places, the actual crazies recognized that the researchers were not insane. My point, and I do have one, is that other IT workers can readily tell who is competent and who is not. The mismanagers I know and have worked for are more interested in who spouts "business speak" and who "manages up", promoting them over the people who actually know what is going on. Dilbert, and thedailywtf.com are documentaries of the IT world. In short, politics is far more important in modern corporate America than actual merit.

"IT" is actually even less useful a term than "engineer".

It ranges between the night operator in the mail printing room of some bank to the ad-sense design team at Google. The ratio in skills, foresight, ability to execute complex things, and so on, is huge. The ratio of compensation is therefore huge.

The odd thing is that past a certain level, things like "be proficient in language Q" are irrelevent. If you have the skills to work on, say, pagerank, or windows, or ios, then 'c' vs 'assembler' vs 'language of the week' are irrelevent. Can you grok multi-thread synchronization across distributed clusters of machines? Can you design a sound new object class? Can you study a network protocol and understand why it is locking up? Do you have the vigilent mind-set required for work on security, encryption, or error recovery?

Chris wrote "The biggest issue I see in recruiting talent is a severe lack of people that understand complex systems." - that's closer to the point.

Is there a shortage now? I don't know. But look at the number of open reqs for high-comp positions at tech firms known to hire in large numbers. Look at the salaries.

Lots of open jobs for "skilled welders" at $18/hr? Probably a wage clearing issue.

Lots of open jobs for "kernel developer" at $90K/yr + benefits + stock? Maybe there's a shortage of people well suited for that position.

'Lots of open jobs for “kernel developer” at $90K/yr + benefits + stock? Maybe there’s a shortage of people well suited for that position.'

But who wants to work as a kernel dev on a crappy, frustrating, spagetti-code kernel that sucks, when they can work on a really awesome one as part of open source?

But who wants to work as a kernel dev on a crappy, frustrating, spagetti-code kernel that sucks, when they can work on a really awesome one as part of open source?

Someone who wants paid?

There's a non-trivial number of really smart people getting paid $90K/year *and* working on open-source projects. Browse a kernel mailing list and increasing amounts of code is being pushed by people employed by IBM, RedHat, Dell, VMWare etc.

Solaris is no longer open source.

But Illumos is. Most of the smart guys from Sun that worked on Solaris have found new positions in the Illumos ecosphere (e.g. Joyent)

Good riddance to Oracle's version of Solaris.

> Lots of open jobs for “kernel developer” at $90K/yr + benefits + stock? Maybe there’s a shortage of people well suited for that positiion

I'm sure there is! Unless by kernel developer you mean "I once had a dream that I changed one line of code in a simple device driver, and even then it didn't compile" , in which case there is indeed a wage-clearing issue :-)

"Lots of jobs for brain surgeon at $90k/year + benefits + stock?" - shortage?!

"Lots of jobs for corporate attorneys specializing in bankruptcy proceedings with respect to multi-billion-dollar cross-border failures such as Lehman Bros - $90k/year + benefits + stock?" - shortage!?

And yes, the answer _is_ shortage ... but how interesting is it?

90K for a kernel developer would be extremely low pay, at least in Silicon Valley/Seattle. That is the kind of pay that you offer to 2nd tier new grads.

Look at what Microsoft pay Software Development Engineers, which is the entry level for engineering at Microsoft. Even Software Development Engineer II tend to make a lot more then that.

Source: http://www.glassdoor.com/Salaries/microsoft-salary-SRCH_KE0,9.htm

> 90K for a kernel developer would be extremely low pat
Exactly! Extremely low. And relative to the cost of getting a degree, it's astonishingly easy to demonstrate to the world that you have what it takes (*if you do*) - be active and effective in open-source Linux or BSD (or other more obscure options as well) development. You can do that part time. The alternative is what - living to get a degree?

So we have $150k/year starter jobs (let's drop this 90k fantasy), with work that many people find find an intellectually engaging (yes, I know, a small minority, but still many of us are geeky and/or aspergery), where you have a clean path to proving your ability pre-hire without sacrificing the rest of your life while doing so.

I'd like to know how that is remotely consistent with the thesis that the cited paper applies to I.T. (Which, by they way, they do not say: their examples are about show business and such which I cannot comment on.)

I'd also like to know how the above fact is consistent with "incentives are too low for Americans", but if we raised them all this world-beating talent would come out of the woodwork!

Are you looking at a different link than the one I see? The average for SDEII there is only 97K. Not that great for someone with 5-10 years experience. And they're outnumbered 3:1 by the SDE I at 88K.

The better salaries in my experience tend to be for consultants -- that's where the ambitious people tend to go. And if you can start your own consultancy, or even just work on your own, you can make $200K+ pretty easily.

I'm not commenting on glassdoor links about one particular company, even if it is Microsoft, and trying to parse its idiosyncratic titles (though I note that the top salary of their whatever-it-means "SDE1" is far greater than the top for "SDE2", so maybe these titles don't mean what they say or the statistics are incomplete (Self-reported stats from the web, categorized in an obscure way, might have problems - who would have thought!)).

We were talking about _kernel developers_, and if you are one such as is usually understood, something would have
to be really VERY constrained or strange about one's situation to make under six figures. That's just not near the market price.

Thinking back to the original article, which yes I have read, if you are a decent software developer still making <100k after 5-10 years[*] that is the signal that you probably are (what the paper calls) "mediocre" and a threat to optimum economic efficiency (as they see it) :-) Myself, though I disagree with the philosophy of this paper if not its conclusions, I'll happily hire top-1% new grads at 1.5 to 2x this rate all day long and be thrilled about how you are keeping the market price down.

[*] I'm a bit uncertain here. Is it really even possibly true that a _good_ software engineer at microsoft is plausibly at 97k average after 10 years? (total comp; so stock included). If so I'd have to agree that something very fishy is going on, likely involving the H1-B process. Though come to think of it, this does explain a lot about MS in recent years!

I agree kernel would be different, I just didn't see much support for the claim SDE II make "a lot more" than 90K. They make a bit more, on average.

I agree that one should expect to make >100K after 5-10 years, as an ambitious developer. My assumption was that MSFT was hiring a lot of drones. I offer Vista as my irrefutable evidence :)

Otoh my wife is a very good J2EE programmer and doesn't make that much, but she came here on an L1 visa about 5 years ago and has only an associate's, so different situation.

TallDave: There is another 15K in stock & bonuses that were not counted as part of the 97K.

Sure, but the comparison was to "Lots of open jobs for “kernel developer” at $90K/yr + benefits + stock?"

SDE II is 1-2 years experience. SDE I is kid fresh out of school.

Still seems low, but (again, read the paper!) I guess Microsoft is in the talent-determination phase with such new hires. And yet, among those SDE1 ("fresh out of school") someone reports an $192k salary. Notwithstanding my skepticism about glassdoor stats, but I totally believe that. For sure, there has to be something quite special about you to get $192k fresh out of school in software engineering, but not like drafted-to-the-NFL special or rare. I must say that $192k for a fresh grad _in a mature and stable company like Microsoft_ moves it from the "well done" to the "impressive!" league.

There are 3x as many SDE1, and 10x as many SDE as seniors -- that would imply 90% of people lose their jobs in three years, assuming similar reporting rates. Surely MSFT retention can't be that bad.

This is glassdoor. As I understand it, the #'s your are seeing (both count and averages) reflect those who have voluntarily signed up to this independent website, and voluntarily provided their position and salary information. It is not the number of such employees in the company nor an unbiased statistical sample thereof.
One would expect newer employees to be more open than those more established. Likewise the unhappy vs the happy. Likewise (since glassdoor is somewhat new in its popularity) the younger vs the older. If you see glassdoor stats that seem true even counter
to these (at least these three, there are more) biases, it's potentially interesting. When you see something interesting that is potentially explainable by the self-reporting bias, it's hard to be excited.

Possibly, but then it could be argued the other way too -- people who make more are more likely to report, people have been there longer are more likely to have posted as a function of time, older workers might use career tools more, etc.

Well, we can't say - the only point is that this is not a statistically random sample so one needs to beware.
The more potential biases you suggest ... whether or not I agree with them ... only further damage the idea that these should be read as population statistics. While I think there are problems throughout, I am especially inclined to think your comparison of respondent ratios (number of those answering at level SDE1 vs those at SDE2) is too compromised to be of particular value to anything and in particular career mortality. I have no strong view on this either way, but this seems a bogus statistical source wrt this specific question.

Oh sure -- it's not even THAT good, because people could be lying, just for fun.

I wouldn't trust the salaries within 10%, but OTOH 3x and 10x ratios probably mean something.

"the nature of the talent revelation process, rather than by an underlying scarcity of talent"

Reminds me of the NBA and Jeremy Lin.

Very nice paper and interesting contrast to Baumol's cost disease hypothesis.

"I argue that high incomes in professions such as entertainment, management, and entrepreneurship, may be explained by the nature of the talent revelation process, rather than by an underlying scarcity of talent."

But in medicine where you do not have these hudge income differentials, this is *not* a likely explanation for the high salaries.

why is there a shortage of talent in the economics profession?

One thing left out of this story... the powers to be at most Big companies (E.G. Management, HR) really don't like IT people aka Geeks. It ranges from "dislike" to "overt hatred". So internally, most companies don't want to invest in their IT people regardless of how good they might be.

That's an odd statement. You should find a company that doesn't suck.

In my experience, if you work at a company that doesn't suck, it will inevitably be purchased by a company which does suck.

I can only speak knowledgeably about web application development, system administration and networking - but in my experience there are two kinds of IT workers:

1) Those that seem to born to the job - able to master the skills, extrapolate and solve problems quickly, master new technology quickly and who absolutely love what they do.
2) The much larger group of "normal" people who are confused and bewildered by the changing technology, the complex problems and all the special cases.

In my experience, what you're seeing, wage-wise is not a place where the older workers are being chased and the young workers ignored - it's that the group 1 employees of any age are being chased, and the rest are ignored. The young group-1 people don't have enough experience to really kick their salaries into high gear yet, but they will.

The people in group 2 are slightly better than mop-water, but not much. No one wants them if they can avoid it, and they can't be trained to be group 1 folks. By the time they are older, they've settled down in a niche and they can't break out of it, because they're not good at self-starting, they don't love what they do, etc.


The group ones are gold. They absolutely love programming, period, and keep up with technology because that's also their hobby.

I just wanted to say that I knew CBBB would be in here, whining and moaning about his inability to find a job. I know I don't deserve a prize for something that obvious.

It's also interesting, as always, to see the self-imposed "beta" take on the world.

I worked in HR for a while, and while there is a lot of form checking that goes into it, it's designed to try to weed out the CBBBs of the world.

Here's a hint: you may be "can do," but you certainly don't come off as "will do." That's just the start, but if it gets you off your couch for a day, it's worth it.

Uhh Actually I was "whining" about how obvious and trivial a conclusion this Sorkin came to. I think it's ridiculous people get paid (probably a lot) to take some integrals and come to common sense conclusions

the nature of the talent revealing process rather than an underlying scarcity of talent

still think its ridiculous?

now do you see why the paper is interesting?

Yeah I do think it's ridiculous because it was always the most obvious thing in the world to me that there was no scarcity of talent. I understand economists don't know anything about the economy so I guess it's interesting to them - I don't understand why these economists are of any value what-so-ever if they think something trivial like this is insightful.

Are you just weeding out CBBBs or creating more CBBBs?


It doesn't need a whole department to detect and avoid the obvious.

Every time a new HR manager is hired a new CBBB is created.

Let's kill those italics. A B.

Well? What's stopping you?

Why are there shortages of affordable beachfront property in Santa Monica? Why are there shortages of jobs that pay seven figures but don't require much work?

Oh, wait, economists have told us for generations that the word "shortage" isn't useful in economics. But, over the last couple of decades, economists have come to assume that the word "shortage" is perfectly appropriate when used from an employer's point of view: e.g., Why are there shortages of 200 IQ employees who will work for minimum wage? Great question!

It doesn't seem like there is an actual "shortage" of IT workers. Perhaps at some low wage there is, but then you should just offer higher salaries.

Hold on... Looks like you guys needed an I/T guy to close that open tag.

Trickier than it looks. italics with an i with an em now let's try... But I'm not optimistic at this point...

See? As an old I/T guy, I'm worthless.

>>> why don’t firms train more workers<<<

Because the concept of "job" is obsolete. Employers don't value employees, they value results. As such they couldn't care less if these results are provided by existing employees, subcontractors, or new employees. Sure, all employers say "employees are our most important asset", but that is more false than true. The true part is they value perhaps 15% of employees who are the key high performers. The concept of "job" in the 20th century was based on some preconceived notion that a job endures if you simply do your job -- that is no longer true. There are no "jobs" anymore.

Some years ago we always had an unwritten premise that the performance of software people varies by an order of magnitude. When you are in the next software engineering meeting, look around, one person will be 10 times more productive than another person in the room -- this is still true. Yet company rules and perhaps employment laws do not allow compensation according to actual performance, the highly productive person will make perhaps 10-20% more than others. Unfortunately companies also value non-technical skills in our population of IT/SW people, like political skills. Often the ones earning more are better at politics than technology. The only way to equitably compensate the top SW people is to promote them until they are in a job they cannot perform. We are at fault as well, not just the employers.

Also some years ago it was the obligation of a "professional" to keep up their skills, not to wait for the employer to create some training program. But at some point software / IT skills became so important, and that employers paid a lot, and then we workers started to feel like elite and privileged workers. I think we got lazy, we didn't invest our own time off the job or on the job (during lunch perhaps?) to keep our skills up to date. And we got mad that they made us work all these long hours on high pressure projects in the absence of "equitable" compensation. We forgot that we were already getting paid very well. And when our stock options didn't fill this perceived compensation gap, we blamed management all the time forgetting that the work was interesting and the compensation we did get was actually pretty good. We are at fault here also. If you spend more time on Facebook than retraining yourself you may be one of the people whining here in the comments.

Don't think "jobs" anymore, think "profession" and "value / results", and specialize.

Thanks for the commercial from the Management class.

Oh, don't worry, I am one of you. Just sharing some lessons learned the hard way. The discussion was too focused on surfacing problems, not trending towards ideas to solve these problems.

True I have had some significant management responsibilities, but honestly I always hired the best people I could find, not those that checked off the most language and technology skills. I would hire a PhD in Biology or Chemistry over a BS in Computer Science or engineering if the person had some exposure to computing and an understanding data, and demonstrated that they could think. The important skills are critical thinking, mathematics, logic, and skepticism. These skill endure through changing technologies. We all need to think/behave like PhDs. Why? Because they are trained to uncover new knowledge, and new problems to solve, and create solutions.

Again, it is not about "a job" or "the work", it is about solutions to problems. We need to show our customers (oops.. slight slip... I meant potential employers!) how we can solve their problems, not how we can fill a job.

It isn't law or rules that keep firms from paying top people an order of magnitude higher than others. It's budgetary constraints: by and large: they can't afford to do that, because their products will not sell if they were priced to include those types of salaries. The market just doesn't want that level of productivity. It's satisfied at the mid (or high-mid) level. Basically the same reason the Concorde was mothballed, and other high cost, over-engineered products do not do well.

Wandering a bit further afield here, but some advice for those wishing to land a STEM job -- be careful what you wish for:

Life in IT is truly a special kind of hell for those who have no interest in, or, as it came to be in my case, an intense loathing of, the work.

I majored in CS, not out of any genuine desire to delve into the intricacies of sorting algorithms (I always found math and science tedious, at best), but because of the starting salaries, along with the assurance that it would always be in demand (You can do anything!). I hated every minute of every class during undergraduate, and I slowly came to detest my peers for what I saw as their immature hobbies, slob attire, social maladroitness, poor hygiene, whatever negatives I could pin onto them (looking back now, I can see I was allowing my bitterness to color my every interaction with them).

Graduated in '09, and got a development job paying 65k a year in 'flyover country'. Loathed it with every fiber of my being. I was most likely a dreaded 'ZMP' worker there, since 99% of my time was spent devising exit strategies (not only from the company, but from the STEM field). Finally jumped ship (now out of the industry altogether) last fall, taking a painful pay cut, but no longer grinding my teeth down to nubs or shouting obscenities during my commute.


You sound like a perfect candidate to consider business school, or law. I have worked with colleagues similar to yourself with undergraduate STEM degree, weren't satisfied, and then went back to law school, or business school, or in one example med school. The combination of CS and another specialty made them highly desirable and very much in demand. This later specialty becomes the focus, not the CS part, but various other disciplines need people who understand the basics of CS and can apply that knowledge in ways that the others can't.

I happen to love STEM, because after some years I still have a longing to figure things out, in a way this need has grown the more I learn. I view it all, including CS, as simply tools to model the real world; CS and SW is not a goal, it is a means to an end. Creating models of the real world lets us understand what is really going on, and therefore make better decisions about what we should change, and how things could be better in the future. And also, if other people see these models, and can understand them, it can be a pretty rewarding thing. Likewise creating things that people like and use (some of mine still being used after many years) is a really good thing.

I share a lot of the sentiments of Matt but, as I've said before, most STEM graduates are going to have lower marks then you're typical humanities grad and therefore are at a huge disadvantage when it comes to applying for Business school or Law School. Maybe Matt has fantastic marks but if he didn't love CS he probably just has decent but not the earth-shattering marks you need to compete with the grade-inflated humanities grads.

I considered law school, but I am very, very, very wary...I have friends who graduated from T-14 schools (I am not sure of their class rank), who are having trouble finding work. Fortunately, the field I am now in is less concerned with credentials than results (this also being a positive aspect of the STEM fields....).


Yeah --- marks, grades, evaluations, performance reviews. These may not be completely under a person's control, and are absolutely not final judgements. What is more important is to do what we love, what turns us on, and we gotta go do it, and then, guess what happens, we do it well.

My neighbor who I've known for many years is a heart surgeon. I really love to talk with him. Not even close to rich, he works at the local community hospital, a pretty humble guy. Goes in to "work" at the "job" at all kinds of hours, he repairs hearts for the elderly, the mature adults, early adults, teens, and children. Nice house, nice cars, (nothing rich here), and his children are really good people.

As in prior posts I read about being able to grok multi-threaded synchronization (BTDT, I am expert, learned my first SW in embedded multitasking systems), I would love to be able to grok security (but don't yet understand it), but also had the pleasure being wired in: coding so fast and accurately for weeks/months I can't believe it came from my fingers (a great experience).

But man, I can't imagine the ability to grok as Rob has experienced while operating on a human heart hundreds of times, maybe more. His hobby is short wave radio, and experimenting with packet radio, where he asks for my help. And then, I look him in the eye, and I feel small -- which is always one of the best experiences in my professional life (get it?)

I am a java programmer and I have come to the belief that we (IT workers) need some kind of guild (not a union) to represent our interests.

Most people don't like to be perceived as computer nerds therefore they will hide talent until compensated at a rate equal to the market rate for the work being done plus a nerd premium.

It is apparently socially acceptable to not know how to perform basic computer functions in front of an audience. For evidence, go to any talk, presentation or seminar where technical difficulties arise -- even among people who consider themselves nerds.

This is similar to the social acceptability of people saying they are bad at math or unable build IKEA furniture in front of an audience. They are either inept or, more likely, signalling the social desirability of those types of skills/occupations.

I once was negotiating with HR for a new hire. They asked me to describe everything I needed in this person. I said, "Familiar with computers and able to learn." From this, they produced a 2 page document that included degree requirements and specific skills (often misspelled, such as SEEQUAL, instead of SQL). The kicker was that I could not fulfill the requirements listed myself.

My conclusion: At least half the problem with finding talent, is caused by the extreme lack of talent to be found in HR.

After reading the paper, I've come to the conclusion that the author only has a weak argument where "talent" is objectively measurable, such as in sports or entertainment. In my experience in IT, managers very rarely have a clue who is talented compared with who is not. And his suggestion to have new workers pay to be hired (page 2, para 2) strikes me as fantastically ignorant. I can already see it being abused to charge workers to sign up, then fire them before they've earned back their hiring fee in wages (this is common in "work from home" scams).

Generally, in IT, talent is grouped into 2 basic pools: junior and senior. The pool of junior talent is constantly filling due to entries from colleges and the self-taught. The senior pool is filled from junior talent getting skill on the job. Most companies and recruiters only want talent from the senior pool, with the silly long lists of required skills (which have been mentioned by others commenting) being a proxy for measuring who is "senior" and not - with the added problem that many employers only want to offer "junior" wages for "senior" talent. This leads to the age-old lament that it is impossible to find affordable staff when one is not willing to pay them what they are worth. And the inability to determine workers' worth adds only to the dilemma. Getting hired in IT is a lot like dating in high school: you are desirable if and only if you already have a date/job, lose that job/date and you suddenly get a severe case of cooties. This cooties contamination spread to enough other professions during this recession that congress passed (was going to pass?) a law banning job adverts that said things like "no unemployed need apply".

After one period of work, those whose talent is revealed to be below a certain threshold level exit the industry, while those above the threshold stay on
This is an awkward way of saying that underperforming workers get fired.

One of the biggest changes that could be made in IT would be to implement some version of the system used at law firms and consulting companies which was developed at Cravath in the early part of the 20th Century. One interesting article that inspired me to research it enough to start the article on Wikipedia was: http://ww.thedailywtf.com/Articles/Up-or-Out-Solving-the-IT-Turnover-Crisis.aspx

Most people don’t like to be perceived as computer nerds therefore they will hide talent until compensated at a rate equal to the market rate for the work being done plus a nerd premium.

The information inefficiency is severe: but the info exists, locked up on co-workers' heads. People do want to help GOOD co-workers, and all that's needed is to get the info out. That shouldn't be too hard with apps for sending thanks, like Propadoo and Laudits; and in general, the LinkedIn network does a good job of reducing inefficiencies.

It is insanely expensive to train these people. It's a craft - it takes decades to get properly good at it. Why the hell would you hire a novice who won't even earn you money for two years?

Comments for this post are closed