Amazon.com has been around for awhile, and it has been successful for awhile, so I find this especially surprising:
------------
Amazon.com, notably, has a median stay with the company of just one year, a figure Ms. Bardaro ascribed to the intense pace of work there. (The study did not include workers in Amazon’s warehouses, where skills and turnover are different.) “We’re based in Seattle, and know a lot of people at Amazon,” she said. “The consensus is that you are run through a gamut there, make money, burn out and leave.”
------------
I would not have thought that this was sustainable, but if this is accepted practice at Amazon, then perhaps this is sustainable. I am somewhat surprised that Amazon finds this productive. But perhaps they get enough out of that 1 year average that it offsets the costs of having to hire new people every year, and the cost of teaching them how the company works.
The little dirty secret of the tech industry is that if you put yourself in the right time at the right place and get the strategic advantage you can make so much money that a lot of mistakes can be hidden.
Warren Buffet says that when the tide goes low you can see who is swimming naked. The opposite is true, when the tide is high you cannot see who is swimming naked, so one should not make too much assumptions about their clothes.
In the late 90's everyone thought that microsoft had found the perfect way to manage programmers and run a software company. All kinds of companies would try to emulate microsoft methods. But in reality, that clever fox Gates guided Microsoft into the perfect strategic position. Once microsoft had to compete in fields that were outside their monopoly the results were not as dominating.
Similarly Amazon is in a great strategic position nowadays, because they have built out the infrastructure that the US uses to shop online. Their stock is flying high, so everyone assumes they must be doing something right. And they undoubtedly are.
But that does not necessarily mean that their practices of managing programmers are optimal or even good. In fact if you look at Amazon's finances you are stricken at how little profit they make. It is unclear why, because their financial statements are very opaque.
But it is entirely possible that they are hemorrhaging money because of the way the hire and fire programmers but the whole thing is hidden by their strategic position. It is not like a company that does software development better can come in and beat Amazon at selling things on the internet.
> The little dirty secret of the tech industry is that if you put yourself in the right time at the right place and get the strategic advantage you can make so much money that a lot of mistakes can be hidden.
I wish I could staple this to the top of every discussion about the management practices of dozens of companies that come up on HN.
When you're floating down a Mississippi made of money, the sheer current of cash will move you along. You require none of the complicated and difficult additional mechanisms and culture to cope with the dangers and difficulties of the deepwater maritime world.
An injudicious observer might conclude that all the guff that comes with facing the open seas is unnecessary -- why, look at this luxurious barge! They have no need for the drills! No need for a captain! Navigation is a wasteful chore!
The problem is that success is not a discrete, directed acyclic graph. It is integral calculus over fuzzy numbers.
And to continue your metaphor, when there's a bit of a cash drought, your steamboat may run aground on a suddenly not-hidden sandbar, or get stove in by a no-longer drowned tree corpse.
> I would not have thought that this was sustainable
Having lived near Amazon & with many friends who have worked there, here's the secret: Amazon pays a significant premium for new grads.
Starting salaries are 20 to 40 percent higher at Amazon than at Microsoft, Google, & Facebook in WA. However, promotions and raises are also much rarer, so if you actually stay 5 years, it probably averages out minus the stressful pace.
If you plan to stay in the big corporate game, the smart move for a new grad is to leverage your one year at Amazon into a 10 to 20% pay bump in a move to Microsoft or Google.
This only suggests that Amazon's practice is not sustainable. They are paying a premium yet getting perpetually inexperienced people who won't stay as they gain experience.
Surely there is more going on here? This doesn't make a whole lot of sense.
Sixty hour work weeks are sustainable for most twenty somethings, not healthy but sustainable. If you pay twenty per cent more than MS for fresh grads but get fifty per cent more work out of them you'll come out ahead. This does require real standards in hiring to ensure your perpetually mildly sleep deprived workers can do glue code in their sleep because a lot of the time they will be. This model also requires a constant stream of fresh grads but at market + 20% that won't be a problem.
The turnover does suggest that Amazon's reputation as a hellish place to work is well earner but it could easily be as profitable long term as Google's much more humane one.
I don't necessarily agree that some of the best college students graduating each year are inexperienced. A lot of them have been working for 10+ years on their own stuff and have interned 3 times or so.
I don't think it is deliberate.From my friends who joined over the past three years, I think the problem is that burnout as well as incredibly dysfunctional teams is the reason people quit after two years or so. They have been trying to retain people by throwing more money at them (both those who join and might be aware through their friend network of the stigma) and those who are already working there.
I guess it is sustainable, in the sense that there are far more new grads each year than new Amazon hires, and Amazon can afford to pay extra (as evidenced by their financials). It's just far from optimal.
What this is about is hiring people, loading them up with blame, then firing them. The tactic allows totally incompetent managers to keep their jobs indefinitely. I am talking about so-called "dev managers" who fundamentally don't understand what a web service is. Has zero to do with software development.
I can understand the burnout. I personally love it, but I don't think I'm the norm. It doesn't look particularly sustainable, at least from a traditional HR overview...But I still think they get enough of the masochistic super-geniuses to keep the company moving along at its sometimes mind-boggling pace.
Perhaps the "intense pace of work" is the organizational equivalent of the myriad situations where the marginal returns of an additional hour of work are negative, but it feels like they should be positive, so we work/hack/study a bit more.
"One reason for this, she said, was a function of skills. “Baby Boomers and Gen Xers tend to know C# and SQL,” she said. C# is a software language, while SQL is a database technology. She added, “Gen Y knows Python, social media, and Hadoop,” which are newer versions of those things."
It's an acceptable simplification for (presumably) a non-technical audience. Rather than write a paragraph or three about 'what is a SQL and why should I care', then another paragraph or three of 'what is a Hadoop, why should I care, and how it compares to SQL', they just say "it's newer". It's kind of like telling my grandmother "Yes, I work in computers". It saves a good 60+ minutes of glassy-eyed explaining.
The strange thing is that if you just say "newer technologies" it's both a simpler sentence and less obviously wrong. (As someone's pointed out already, it's still wrong because Python's comparatively older.)
I'm not sure this 'tech journalist' knows enough about the field to make any more than a blatantly sensationalist story.
Out of the companies mentioned as having younger workers (Epic Games, Facebook, Zynga, Google, AOL, Blizzard Entertainment, InfoSys, and Monster.com), only two -- Facebook and Google are what most of the younger CS grads from my school would consider worth joining (not even Facebook so much; I know people declined internships at Facebook to do something more exciting).
"The best companies in techlology" (as determined by what, stock value?) aren't exactly the watermark which you judge the entire industry by. There can be a lot of other companies who are recruiting a good mix of older/more experienced, and younger/more eager workers.
I'm not sure if he's implicitly comparing it to the journalism industry in which (presumably) age counts for a lot more.
Agreed. I avoid Java whenever possible for C# or Scala. This, after many college courses that required the use of Java and later on projects for clients. C# at least lets one use more modern programming practices than what Java 7 (or below) allows.
How many people were entering tech 20 years ago vs now though? Could it be attributed to just more people having access to computers which means more people entering the field than before?
I'm surprised to see Epic and Blizzard there. I thought making triple A titles was one of the most challenging things you can do requiring years and years of experience. On the other hand I can understand why gaming companies are filled with youngsters.
The highly technical work has got commoditised and outsourced to very specialised professionals ("engines").
Making a game in the 80s and 90s meant doing basically everything. In the 70s, even the hardware to run it. Now it's increasingly using tools and hacking some customisation.
The spectrum from making a game, a mod and using gamemaker or something of the sort is getting more diffuse.
Do the engineering professions (civil, chemical, etc.) have the same issues? There are a lot of parallels between the type of work the programmers and engineers do. So, what is the difference?
Other engineering disciplines tend to evolve at a more measured pace and have a stronger institutional memory. They also tend to have structured pathways for taking a fresh grad and moulding them into thoughtful, cautious masters of the problem and solution domains.
In our field, technologies largely align on generations. Older coders have deep experience with tech X, throwing that away seems dumb. So they don't follow up on tech Y.
Meanwhile, younger coders, frustrated with the accumulated cruft of X, seize the freshness of tech Y and become masters at that instead.
One day, along comes tech Z ...
This cycle has already appeared in Java or Perl -> PHP -> RoR -> Node.
Having now gone through several cycles, I partially agree with your view on the tech side. But, I just do not believe things evolve at a slower pace in other fields. Especially the fields that have used computers to enhance their efforts.
How have the CAD and CAM tools transformed the classic structured mentoring process? How has more accurate software simulation changed the view of "experience" in older workers?
Civil and structural engineers don't periodically reinvent mudbrick and then go about proclaiming it as a revolution that will sweep non-agile muck like concrete away.
Well, at least in the case of electrical and mechanical engineering, no company can survive burning through grads so fast. For one thing, the number of engineers available is more limited, because of the (relatively) reduced sexiness of those fields. Also, if your product has to work for 10, 20, maybe 30 years, you have quite an incentive to retain the people that know it in detail in order to cope with the inevitable non-trivial long term support issues.
The really sad part of that statement is that Hadoop is a storage and processing backend, and you can run SQL (like, sortof, not really) languages on top of it, like Hive, for instance.
Apple is also mostly older people (and at least from my informal sampling, a larger-than-normal-even-for-silicon-valley number of non-heterosexual people).
------------
Amazon.com, notably, has a median stay with the company of just one year, a figure Ms. Bardaro ascribed to the intense pace of work there. (The study did not include workers in Amazon’s warehouses, where skills and turnover are different.) “We’re based in Seattle, and know a lot of people at Amazon,” she said. “The consensus is that you are run through a gamut there, make money, burn out and leave.”
------------
I would not have thought that this was sustainable, but if this is accepted practice at Amazon, then perhaps this is sustainable. I am somewhat surprised that Amazon finds this productive. But perhaps they get enough out of that 1 year average that it offsets the costs of having to hire new people every year, and the cost of teaching them how the company works.