Hacker News Reader

There is always a "mismatch" between wants and ability to pay. Tech business is not special. Purchasing labor services is subject to the same supply and demand problems as any other product. It's simply market fundamentals that affect all economic behavior.

If you want a new Lexus, but can only afford a used Corolla, that does not mean there's a "Lexus shortage" ... nor is there a "car shortage". If only a Lexus will do, then the market just doesn't happen.

> The skills shortage issue, in my experience is the inability to find someone with the specific skills and talents for a given job. It is real, as anyone who's tried to hire in certain areas can testify.

You know, you could always just hire someone with potential and put him through a training program to teach him the "specific skills" needed, or just let him learn them on-the-job.

But of course, employers are no longer willing to invest in their employees* , so we have a "skills shortage."

* They used to. It's my understanding that the first few generations of business applications programmers were taught programming from scratch by their employers.

This is my biggest gripe with the modern technology industry. We've gone from "you have the background and skills to do the work - here's how we do what we do" to "You don't check these thirty boxes specific to our company?! No job for you!" It's getting rather absurd, especially when you're job hunting and the recruiters give you the "you have an impressive background and could do this work, BUT you don't have <exact ten things> ergo we'll pass." Very few people take the route of finding talent then training them up. Companies willing to take the time to train their people and focus on finding raw talent they can mold will make a killing. Not only will the people be trained exactly how you want them, but I'd argue they'd be cheaper and you could tie compensation increases to hitting defined mile marks in the training program or demonstrating competencies.

From the flip side, I see the company concern - kind of. If I bring in raw talent and train it, I've sunk $X into the talent and if they leave shortly thereafter, I'm out $X with no benefit. The counterpoint to that argument, I think are signing bonuses and defined time contracts. In exchange for us training you to do this job and completing the program, you get a signing bonus of $X at the start, $Y at the end, and sign a 2 year agreement to work here. If you leave beforehand you agree to payback the $X+Y signing bonus (unless laid off by the company).

There are ways to address the "skills shortage". The problem is, no one wants to take the time and the cost to do it.

>look for talent wherever in the world it is available without having to open up offices there

It's not. Most large companies already have offices over there and could hire those people tomorrow if they wanted to. They bring people into the US with their families to have power over them. A manager without power over skilled people is useless in accomplishing business goals that require those skills. They also are always trying to accomplish personal goals like getting their friends promoted, getting more power generally at the cost of technical people. These types of goals tend to make technical people walk away despite any salary level, hence the need for H1B. Except for body shops, H1B is about power not money or available workforce.

Frankly your vision is limited to the outsourcing industry. Would it surprise you to learn that many many big companies (GE, Msft, GOOG, FB, pharmas, etc.) that form a large percentage of the US economy also hire people with specific skills (say, 3D computer vision or expertise in pharma processes) who by being educated in the US universities are some of the best in the world? No manager in these companies is thinking about the leverage you're talking about when the person's skill can be a do or die for entire product lines.

I'm speaking from first hand conversations with H1B's who go on about how their company is holding this or that over their head such as green card sponsorship. You are completely ignored my point that those employers can hire those people in their countries as those employers that you named have offices in every major city around the world. They have big offices with thousands of employees. Your statement about companies wanting to hire specialized skills is pointless. Everybody agrees with that. Your statement about managers not wanting leverage is just not true. What else do you think it means to want to hire to someone other than to have leverage over a person. I made the point that employers want undue leverage when money is not enough and that is why they use H1B.

The parent's description matches exactly what I know from people on H1Bs in the fashion industry, whose labor market conditions have some parallels with tech. I don't think it's limited to outsourcing companies. Long hours, death march deadlines, and generally higher-than-average salaries result in high turnover in fashion as they do in tech. Employers turn to H1Bs to lock-in a workforce, as, excepting large companies, few will expend the extra time and money (legal services and filing fees) to transfer a visa when they can simply hire away an employee from a competitor. The lack of labor mobility with the H1B is a boon to industries that would rather not reform the working conditions or improve the salaries that causes top talent to move around.

Another factor is location. My company acquired a startup in a tier two city in US, and have the toughest time in getting people there whenever they want to fill up a position. At the same time, there isn't much of that problem in other offices.

Seems like a simple problem to solve - raise the wage offered in the tier two city. If the benefits of moving to a less desirable location for work outweigh the drawbacks, people will do it. Your company probably just doesn't want to pay people enough to move there.

> I've never seen any high skilled jobs going to H1Bs just because it is cheaper.

Are you familiar with the Walt Disney Company.

I'm using the definition of "high-skilled" used by the article, which clearly claims Disney used H1B for routine jobs.

Which specific skills and talents are in shortage?

I'll start. 1. Deep learning, Neural networks, anything to do with AI. 2. Machine Vision, 3. Genomic Science (CRSPR etc. 4. Robotics/Mechatronics... and that's just a start.

And H1Bs are used primarily for hiring people in those fields?

In the current structure these are the kinds of jobs that are getting harder to fill with H1B shortages. I personally know several experts who couldn't get O1 (and some who did) and had to leave the country after losing out in the lottery.

I have H-1B coworkers that are not particularly highly skilled or in those fields.

Ok

> 3. Genomic Science (CRSPR etc.

Did you read the article?

> The mismatch is especially stark in the biomedical field. There, according to a 2014 paper by experts from UC San Francisco, Harvard and Princeton, "the training pipeline produces more scientists than relevant positions in academia, government, and the private sector are capable of absorbing."

Genomic Science is not the same as "the biomedical field." I'm not familiar with the job market in that particular field, but nothing you learn from reading the article gainsays the idea that it's hard to hire people to do genomic stuff.

If only a PhD credential was all that you needed to get a job! Seriously, that is one of the reasons the "shortage" argument breaks down. You want the best graduates and 1) they aren't necessarily always US citizens, 2) not all people with degrees are desirable for every open position

Wouldn't raising the price of labor solve this problem?

1 and 2 are simply wrong.

If you demand someone with a phd from a top 10 (stanford/cmu, berkeley, michigan, toronto, ny, columbia, UW, etc) then yes, I agree -- but there's probably fewer than 200 of those minted per year. If you realize a cmu ML phd holder isn't going to your startup for $125k + lottery tickets and is instead going to a public company for $200k + $100k in RSUs and are realistic about what you can hire, there are plenty of us with track records of delivering products and making our employers money.

That's from my perspective both as hiring manager and an employee. I've been willing to hire both non-phds and outside of the aforementioned schools and really not had much trouble.

Not sure where you are based, I am an MIT alum and founder of a company in Machine Vision + Deep learning. I have posted job postings for Deep learning engineer on Angelist, every forum I could think of, including attending AI meetups here in Boston, and...and guess how many responses I got? zero, zippo, zilch....

p.s. BTW we are more open to Masters/Bachelor's (less researchy) than Phds.

Maybe you could hire promising people and let them learn those skills, especially if you want people who are "less researchy."

I'm not aware of a huge pool of "Machine Vision + Deep learning" people who are trapped in India because of Wipro using up all the H1Bs on J2EE programmers.

I am not hiring from India, my intent to hire was international students. I am with the proponents of H1b reform to prevent companies like Wipro & TCS using them for outsourcing/offshoring. But to say "there is no skills shortage" is at best laughable.

> I am not hiring from India, my intent to hire was international students.

> But to say "there is no skills shortage" is at best laughable.

What is stopping you from filling that shortage by training people to have the skills you need? You're trying to hire students forchristsake.

You once again missed my point, the students i am looking to hire are well versed in AI tech, know a lot about the field, and can hit the ground running (vs some entitled prick who thinks he deserves the job). Broad-brushing I know, but only as a counterpoint to your own.

> You once again missed my point

No, I didn't. I just think your hiring expectations are unreasonable and not at all indicative of any kind of real "skills shortage."

Agree with you. There is definitely no shortage....of entitled, arrogant, lazy, whiny, whingey armchair generals like you.

Instead of sharing your infinite wisdom of who I should hire, how I should train them, and what have you...why not start your own company? Hire all this talent pool you claim exists, and show us how it can be done. Be the change you want. But that would involve moving your fat a$$ and might be too much exertion for your royal laziness

Unless you have started up a business, and done something worthwhile, you can go shove your advice up your you-know-what.

> bobosha 10 hours ago | parent

> Agree with you. There is definitely no shortage....of entitled, arrogant, lazy, whiny, whingey armchair generals like you.

> Instead of sharing your infinite wisdom of who I should hire, how I should train them, and what have you...why not start your own company? Hire all this talent pool you claim exists, and show us how it can be done. Be the change you want. But that would involve moving your fat a$$ and might be too much exertion for your royal laziness

> Unless you have started up a business, and done something worthwhile, you can go shove your advice up your you-know-what.

Hmm. I wonder if your attitude has any connection to your hiring problem. I certainly wouldn't want to work for someone who acts as immaturely as you are acting now.

BTW, I do hire people and I do train them/give them time and guidance to pick up the skills I need them to have, especially if they're less experienced students. I think it's unreasonable to expect that the perfect candidate will just fall out of a university fully formed and into my lap, so I deal with it and do something about it. Apparently you do not, and instead jump to conclusions and hurl put-downs.

Not everything can be acquired by training. You are assuming an unrealistic utopia where every skill can be learned. Unfortunately most CS and EE students cannot be trained in AI/ML. Only the subset with strong and deep backgrounds in linear algebra, etc. can be. If everyone can be trained for everything, what is the point of specialization?

> Not everything can be acquired by training. You are assuming an unrealistic utopia where every skill can be learned.

That's true, but only to a degree. There is a kernel of inherent talent/ability behind many skills, but pretty much everything else besides that can be learned.

So, assuming you're correct, you identify the "CS and EE" students that can be trained in AI/ML and you train them, rather than complaining about a "skills shortage" (usually meaning something like the US is not providing fully trained bodies at cheap prices with the cutting-edge, special-snowflake skillset the complainer desires).

"Training" could mean paying for/subsidizing an advanced degree, which good employers have been known to do.

> If everyone can be trained for everything, what is the point of specialization?

I'm not saying something like everyone can learn any job with a 4-week training course. The point of specialization is that no one has the time to learn every skill, so people focus.

>I'm not aware of a huge pool of <blah>

What's the snark about here?

There are lots of Indian grad students in CS and EE departments across the US. Modulo OPT, they're at the mercy of H1B lotteries. I know of PhDs from EU universities who, without the luxury of OPT, haven't been able to convert successful interviews and skill-matches into actual employment because of H1B's arbitrariness. Many of my friends have had to resort to O1 with varying success after excellent PhD dissertations because the failed to make the cut in H1B. Some had to return to India simply because they failed at both H1B and at meeting the onerous requirements of O1.

Did you try hiring from the Brain and Cognitive Sciences department?

I suspect this is a problem: it's hard to articulate the level of skill and competency of what we're trying to hire for. How does one describe a good engineer vs. a subpar one, when the only requirement listed is "knows Java, can code fizzbuzz bug free in 1 function less than 10 lines and in less than 5 minutes?" Something to that effect...