Hacker News Reader

> btw why is he looking through the lens if the microscope is plugged into a monitor? Seems redundant...

Actually, this is pretty common, especially if you are doing fluorescence microscopy (complete derail from your main point, sorry):

- The optics leading to the eye-piece is usually superior to the quality of the camera. You'll have a cleaner, crisper picture looking through the eyepiece

- Most cameras tend to have a refresh rate which is laggy.

- Eyepiece optics tend to have a wider field of view, making it a lot easier to find interesting areas to focus on

- Especially for fluorescence, using the eyepiece helps block out background light, letting you focus on faint details. Some details are just really hard to make out on the computer screen, without fiddling with contrast/brightness settings

- You may have the computer settings at a particular level where "browsing" the field is impossible. E.g. your "picture taking mode" is set to long exposures, or very quick exposures with high contrast. These settings make it difficult to actually use the computer for anything except taking the picture.

- You are taking pictures in fluorescence, but selecting/focusing with bright-field.

I've done this sort of microscopy and can verify everything polyfractal says.

EDIT: sadly, I can also verify that there are no jobs in science :(

> There is massive, and growing, demand in all sectors for software developers, and our universities aren't graduating them faster than demand growth, and as a result software dev salaries are rising (or at least not falling).

This is a direct contradiction to what's claimed in the article, namely that dev salaries are not rising to the extent that would be expected if there were a true shortage.

Because they're not rising. Or rather more accurately, because only one segment of the industry is rising while the rest stay stagnant.

The offers being thrown around by Facebook, Google, et al, aren't anything to scoff at, and as another poster pointed out already, have gone up substantially in a just a few short years (that 50% number is accurate for new grads, it's higher for those with experience).

But the bulk of programmers in the country don't work at Facebook, or Google, or Microsoft. Nor are they forging the bleeding edge future of consumer/enterprise tech. They're doing the unglamorous, not often discussed IT grunt work in the back office of some corporate headquarters.

Their salaries are stagnant, and they are stagnant regardless of H-1B numbers specifically because the supply-demand curve in that leve of the industry sucks. And because our statistics put the $45K VB6 coder in the same category as the $200K+ Google engineer, the rise in one segment of the industry is largely hidden.

There is a danger to making decisions based on averages.

It's strange to me how every time the tech immigration debate comes up Facebook, Google, Microsoft, and co. are beaten to a pulp. They're the ones actually importing people at above market rates, they're the ones handing out $110K offers to new grads and $150-200K offers to even intermediate level coders. They're not the boogeyman under your bed.

Whereas actual exploitative employers like Infosys are barely mentioned. In fact, none of the infamous programming sweatshops are even named in this article. How is that the actual bad actors are facing none of the media scrutiny while the only good (or neutral) actors get all the blame?

If this is the case, it seems that Facebook, Google and Microsoft should be campaigning to have the H1-B quota allocated to whoever is willing to pay the highest salary. That seems like it would be much easier to sell to Congress and it would have the nice side effect of pricing sweatshops like Infosys out of the H1-B market.

That's an interesting point. So we can differentiate between top-talent coders who are in high demand from top companies, and the bulk of programmers working for far less at companies like Infosys, Wipro, etc.

But who are Facebook, Google, Microsoft (not to mention, Infosys, Wipro, etc.) trying to hire through the H-1B program? Top talent? Average coders? Why doesn't Facebook, Google, Microsoft, etc., first recruit from the masses of underpaid Infosys developers? This would force the "bad actors" to improve as well.

Because the masses of underpaid Infosys developers are really bad. I used to work at a name-brand, large corporation known for hiring top-end talent, and did a lot of interviewing while there.

Inevitably the people who are looking to jump ship from Infosys or other low-end consultancy are woefully underqualified for even entry-level positions. I've actually interviewed people from these shops who cannot write Hello World in a language of their own choosing, I've also had people who didn't know what a tree data structure was. I'm unsure how Infosys et al can farm these guys out for such high hourly rates, it's surely criminal.

Sniping from Infosys and Wipro isn't worth it for top companies because there's hardly anyone there worth sniping. They do try sometimes though, since inevitably there is a diamond in there just waiting to get the hell out of dodge. There is effectively zero intersection between the types of developers Google wants, and the types of developers Infosys wants.

As for who Google/Microsoft/etc are trying to hire... I wouldn't call it "top talent" exactly. The H-1B pool in these companies isn't a parade of who's whos of the software industry, though some of them are certainly top-1% coders. It's more like "very good coders" - people who know the tech, learn fast, have solid fundamentals, and get shit done - which is more than can be said for the majority of the talent pool. Simply eyeballing it, I'd say that Facebook/Google/etc are targeting the top 5-10 percentile of coders.

Back in days when I was looking for a job, Infosys was the dream company to work for in India for a lot of top talent. They would hire only the toppers from top 10 colleges in India and that too with a proper puzzles test like all top US companies. So they hire people based only on academic credentials and then train them for 6 months in a "Infosys University".

I think the perception in US about the unqualified Indian engineers is because you might have seen engineers mostly in their budding years(1-3 years). As with everything else, those people do learn a thing or two as they gain more experience.

I came from India(never worked in Big Bodyshop like Infosys) and have been in US for 5 years now. I have seen really a lot of really dumb people and some really good people on both countries.

Check out the puzzles book that Infosys uses to hire people in India. http://books.google.com/books?id=F-idu9oxkdgC&printsec=f...

It's also interesting that everyone agrees that the only thing in short supply is the highest quality programmers, but immigration policy is focused on increasing the total number of immigrants and not just the highest quality ones. Under our current strategy, we'll get a similarly small proportion of top quality programmers along with 97% who are basically there to keep down operating costs for MSFT (or Infosys, Wipro etc - arguably "bad actors" from the perspective of how Silicon Valley likes to do business).

If we were serious about alleviating the shortage of top quality hires, we'd have a way to increase the quota of "the best" technical talent and not the current blanket approach, which really favors major employers trying to save costs. Maybe we let google design a technical test for anyone who wants to work here in a technical role. :)

I think Facebook is sincere in their desire for quality talent, but since their CEO happens to be one of the youngest billionaires in the world, who has only worked at one company which he founded as a pup, he is equally inexperienced in "the real world" (eg, how many other companies operate and what they use H1B for) and probably a very poor choice for a leader on immigration reform.

Yes, they're trying to hire top talent through the H-1B program. Immigration status is orthogonal to talent.

That's the reason they're generally not trying to recruit from the masses of underpaid infosys developers.

I don't know about Infosys, but we have a TCS (Tata Consultancy Services) branch here in Uruguay, and it is very common for junior developers to do a few years there, and then quickly realize they've "outgrown" the company and get recruited somewhere else or start their own company - and those "child" companies recruit the cream of the remaining crop, leaving the worse/less motivated employees at the mother company.

TCS does try to keep some of the better ones around, but from the outside I cannot know how succesful they are.

And that is the reason why there is a mass of underpaid talents.

It's definitely possible that the blue-chip companies are missing an opportunity to develop talent out of that ecosystem, but it's also not their responsibility to do so. Most of us who are senior engineers at whichever high-flying company were, at one point, just working on some boring-ass webapp.

To throw some numbers at this question about salaries. Here is the Standard Offer for a college hire Level 59 Microsoft SDE (Developer) to Seattle over the years as I've observed them.

  2005-2006: 70k
  2007-2008: 80k
  2009-2010: 85/90k (there was a macro recession here)
  2011-2012: 100k/105k
  2013: 105k--110k (tbd)

In 6 years that's a 50% increase. What other industry has as much salary growth for front-line or entry-level people? I believe in a shortage. The most likely cause is this VC-infused hiring bubble going on that may burst and put a lot of people back on the market.

What's tragic is that Microsoft people who started in 2006 and had 70% careers are often making similar money as the entry-level folks. Combine that with a toxic environment and stagnant products, and it's pretty easy to understand why great people leave the company. It's also pretty easy to understand why Microsoft wants to import H1Bs that often are terrified of getting deported back and will do anything to keep their jobs.

> What other industry has as much salary growth for front-line or entry-level people? I believe in a shortage.

From 2005 to 2008, first year analyst/associate salaries at Wall Street banks, consulting firms, and law firms jumped 30% or more. In neither case was there a real shortage of people across the industry. Instead, it was just a matter of a few top firms bidding up salaries for the top young graduates in response to a temporary boom. Certainly nobody was calling for increased immigration to address any perceived "talent shortage."

This is exactly what is happening in Silicon Valley right now. You're seeing a spike in salaries for the few top people at the few top employers as the result of a temporary boom in the sector. You're not seeing a long-term trend of the industry needing more CS graduates than our domestic institutions can provide.

Two things that make me suspect this information is anecdotal rather than from a hiring manager (though I could be wrong):

1) I left MSFT around 2006, but the offers I made to (admittedly, mostly blue chip) candidates in 2002-2004 were closer to 80k and not even at the top end of the L59 level band. I'm not sure those early year numbers are quite spot on.

2) The initial offers are set for the middle of the L59 compensation range/band. As those bands went up, so did all of the others - I guarantee the L59 band is not above the L60 band. If you got hired in 2006, you're at least L62 by six years later (or you would have been managed out, in any non-disfunctional org). Even if you'd been camping and never made L63 (which is a big bump), you would get raises as the midpoint for the band moved over the years. Otherwise, it raises very serious issues with your group's HR generalist about why you have a L62 who's being paid outside (below or above) the acceptable range around the salary band and the manager would have been encouraged to either motivate or manage out the employee, if they really didn't deserve raises to take them back into the correct compensation range.

Microsoft is disfunctional about many things, but the HR generalists were pretty good at tracking and running numbers and asking very hard questions about people who were over/under-compensated or stagnating at a low level (sub-L63).

> I left MSFT around 2006

The company changed significantly since then. Bill left. There was a layoff (well, a couple, but the first was a black swan nonetheless). There is a new review system and new rules about comp.

It's not the same place you remember. I really liked the MSFT around 2006, too :)

I don't know what all that Microsoft-specific talk means, but you can confirm his numbers by checking the H1-B wage data from MS. I don't recall what the numbers are, but I recall it being fairly close to those supplied in recent years.

I was hired by Microsoft as a fresh grad in 2005 for 84000 year. 84000 in 2005 according to CPI calculator (http://data.bls.gov/cgi-bin/cpicalc.pl) is $100,117.42 today. Add to that that two years ago Microsoft reduced stock grants and put more money into base salary and actual total compensation increase for fresh grads is negligible.

I think Microsoft is sort of a special case. Fifteen years ago they could attract the best and the brightest with below-market salaries, both because they dominated the software and computer industries (so your stock options were worth a lot, and your work would be on a computer in every office and every home), and because it was a fantastic place to work.

Yeah, I don't have data on hand to support this, but I also doubt that software dev salaries are really rising fast enough to indicate an actual shortage, or at least a new or growing shortage.

I think the problem employers see is not that software dev salaries are rising, just that they are high compared to other job descriptions, and have been for years. Starting salaries for an entry-level software developer (with a bachelor's in CS) are roughly double the starting salaries for non-programmer entry-level positions that require a bachelor's degree--and employers don't like that.

True. It would be interesting if they could break out salary growth for top developers vs. average developers. The common sentiment here seems to be that there's a shortage of highly talented developers. If this is true, then their salaries may in fact be rising at a higher pace.

"Starting salaries for an entry-level software developer (with a bachelor's in CS) are roughly double the starting salaries for non-programmer entry-level positions that require a bachelor's degree--and employers don't like that."

That's also true, but it's also interesting to consider from the perspective of a talented student considering CS. If he believes everything he reads, (a) he'll start with a decent salary but probably hit a glass ceiling relatively quickly (a much lower ceiling than, say, a doctor or management consultant or trader), (b) his skill set is in constant danger of becoming outdated due to the unusually high pace of skill turnover, at which point he'll be replaced by an H-1B holder or new CS grad, (c) he'll be expected to move to Silicon Valley (or NYC, apparently) where the cost of living will negate his higher salary. If any of these things actually come to pass, the relatively high starting salary he made the first few years out of school won't be much consolation.

That's a good point. I have difficulty reconciling two opposing feelings: I feel like the US is not producing enough developers/engineers, and I feel like the salaries do not reflect the shortage.

I have been working as a developer (and also an analyst) in the DC area for 6 years. The talent pool around here is very shallow. The companies for which I have worked are constantly looking to hire programmers. It is very difficult to find qualified applicants. Roughly half of developers at my current company are on H1-B visas. I think H1-B is both a form of indentured servitude and somewhat necessary (at least in the DC area). I wish there were a better way to employ immigrants, so that they could more easily job-hop or go out on their own.

OTOH, I think the salaries have not reflected this shortage. Maybe I am just a poor negotiator.

Perhaps Boston, Austin and other hot beds are different (i.e. deeper talent pools, more opportunistic employers). But I see my employers struggle to find developers, and poor salaries blamed on the economy.

If the salaries haven't reflected the shortage, it's not a shortage. It's just a bunch of people unsuccessfully trying to buy something below market price.

> But programming is not STEM, and these jobs are not scientific research jobs.

STEM stands for 'Science Technology Engineering Mathematics'. Programming is part of the middle 'TE' bit plus a little of the 'M'.

That's a problem with the definition of "STEM." It's a categorical grouping that doesn't reflect any real relationship. Building web applications is not even remotely related to writing research papers about microbes or subatomic particles. It's a completely different job market. When people say "there is a shortage of STEM graduates" they usually just mean CS graduates. But the journalists writing this are too non-technical to know the difference, so they try to refute this by pointing to statistics that show the US has plenty of STEM graduates. Too bad they're mostly biologists and not programmers!

It's obvious companies want to increase H1-B for hiring top talent at lower salaries.

An amazing programmer in India or China probably makes 1/10th of what he would make in America, but they don't know this. They'll be recruited to America on H1-B but they won't contend their salary, they'll be happy making 2x or 3x their previous salary, even if it's 45k/year. Their ability to negotiate salary will be further hindered by a strong possibility of poor-English skills, therefore inability to communicate well.

> why is he looking through the lens if the microscope is plugged into a monitor?

Must be one of those newfangled stereographic retina-resolution microscopes.

It appears to be a binocular microscope, but not a stereo microscope. A stereo microscope has two separate optical paths to the specimen, one for each eye.

A non-stereo binocular microscope has one optical path from the specimen to the eyepiece assembly, and then uses a prism to split that so that it can be sent two eyepieces. This increases comfort and reduces eyestrain compared to a single eyepiece microscope.

I feel like the key difference between STEM and software jobs is that STEM jobs are typically niche jobs, but software applies to all fields. Every company can benefit from a good software engineer so there's huge demand for it and I don't see it ending any time soon.