分享今年看到的最好的文章# DataSciences - 数据科学
c*z
1 楼
http://www.quora.com/How-long-will-coders-salaries-remain-as-hi
与数据科学的关系不大，与数据科学家的关系比较大。文章分析了一直困扰我的一个问
题：为什么老板都是外行？答案是：我们被外行殖民了。
This is an odd question, because software engineers' salaries aren't that
high. I wrote a blog post on this, here: Software engineer salaries aren't
inflated-- at least, not for the 99%. Relatedly, I wrote a Quora answer on
the topic: see Michael O. Church's answer to Why do software engineers make
so much?
That said, I certainly have no illusions about the impossibility of an
industry-wide salary drop. It could happen. I think that, for the less
skilled programmers currently able to make $120,000 per year on the strength
of their degrees alone, it will. They're going to be hit harder than the
genuinely capable engineers, who might see a 20% drop at worst. Of course,
there's a duality to these sorts of things. Where there are pain and
obscurity, there's also opportunity. Google accelerated in the dead times of
2002-4 and Facebook was founded in 2004, when society at large had little
faith in startups. (I knew people who turned down 2004-era offers from
Facebook to take shitty, 100-hour-week, "analyst" positions at Goldman Sachs
. Startups were way out of style until about 2007.) Now, startups are "cool"
again, and a bunch of dogshit is getting funded and it's harder than ever
for genuine talent to be noticed. So there is bad in good times, just as
there is good in the bad ones. My point is that this possible (and not
certain) impending salary drop isn't necessarily a bad thing, if it hits the
underqualified the hardest and burns out the underqualified executives
first. I could survive a drop to "only" $120,000 per year; these
narcissistic poseurs, currently making a quarter million in their VP/NTWTFK
(Non-Technical Who-The-Fuck Knows) positions, couldn't.
If salaries crash, we're going to see a decline of interest in programming,
and here's what's going to happen. The less-capable programmers are probably
going to be flushed out of the system. The capable ones like me will be
hurting, but it'll be a small salary dip: probably more like stagnation than
a fall. Socially astute people will realize their improved leverage with
the thinning crowd and grab more opportunities. When no one's making money,
you want to optimize for interesting work and for status; when the market
turns favorable again, you have better odds of having the "overnight success
that takes 8 years" and hitting it big. If you want to gain a couple levels
in "data science" or management, or to cross the chasm and become a VC, now
's not the best time to ask, because the field's a bit congested with people
willing to accept maltreatment because they expect to be dot-com
millionaires or startup CEOs inside of four years. If the market crashes
tomorrow, then two years from now will be a great time (if you're good) to
start pushing forward on the autonomy/status axis. It'll be hard as hell to
get a raise, but easier to get the title of "Chief Architect", or to get
introductions to VCs (who won't be considered "cool" or inaccessible, post-
crash). Six years later, when times are good again, you can win.
I don't see "new people entering the field" as a major issue, because if you
give programmers autonomy and resources they tend to advance the state of
society, and that creates jobs. It's not about the number of people, but the
conditions they're willing to accept. In fact, what keeps our salaries
mediocre and our status and autonomy so low is that we're terrible at
organizing around our interests. We should have a union or professional
guild, and we don't. We should have private talent agents, and be organized
enough that we can afford to pay them. We should have an employment field
where binding mandatory arbitration, fast firing without severance,
aggressive non-competes and "we own your side projects" clauses don't exist,
and where we have legal representation at employer expense in the event
that our rights are violated. If a history major with stronger social skills
, who discovered a passion for technology in her mid-30s, comes in to our
fold, I think that's great. We need more diversity of thought and
perspective. We need more socially skilled people who identify as
technologists and engineers so we can lead from within instead of being led
by the clueless, short-sighted McKinsey culture that is currently colonizing
us (see: The Damaso Effect).
Now, the H1-B program has depressed wages, and that has a lot of people
scared. The issue isn't classical wage depression, though. Once you bring
someone to the Valley, his wage expectations are the same as anyone else's
after 5 years. Someone who's ambitious enough to come to this country isn't
going to be happy to be mistreated and underpaid just because he's from a
poor country. The major failing of the H1-B program is that it creates a
pool of essentially captive labor, because losing your job means you'll
probably have to leave the country, and quickly, and international moves
without a relocation or repatriation package (and startups don't pay for
those, because most of them are shitty) are terrifying. The H1-B program was
designed to bring in high-talent outliers, and instead, it's being abused
to create a pool of slave labor. Its effect on wages at the lower end is
small (H1-B workers still get middle-class salaries; the program isn't
nearly as abusive as, say, Japan's programs for Southeast Asian tech workers
) but the program's general effect on programmer autonomy and status is
devastating (and that hurts our salaries, later on). I have no problem with
the number of people brought in by this program (in fact, there aren't many
H1-B slots) but, if we're going to have that, then admission to this country
has to be unconditional with regard to employment; once you have it, you're
good for a certain number of years no matter how often you change jobs.
Fixing the H1-B program should be a high priority. And that's, again,
something that probably won't happen until we (as technologists) bring a
more diverse set of people into our fold.
Okay, with all that said, time to answer the question: I don't know. Knowing
that there is a bubble is easy (we're in one) but separating the genuine
changes from the bubble effects is very hard. (See: the 1990s bubble. The
Internet really was a BFD.) Predicting when a bubble will bust and how is
probably impossible. Software engineer salaries aren't that high, so they
could credibly increase or decrease in the next five years. I wouldn't worry
about it, because I don't see them changing by enough to sway career
decisions-- at least, not if you're capable and passionate. Whether the
economic tide goes for or against our industry, we need to get much better
at organizing around our interests. If times go well and there's more
prosperity, we deserve to share in it, instead of letting well-connected,
pedigreed McKinsey assholes take everything. If they go poorly, we want to
be protected. Either way, we need to get better-organized, and there's no
time to waste.
与数据科学的关系不大，与数据科学家的关系比较大。文章分析了一直困扰我的一个问
题：为什么老板都是外行？答案是：我们被外行殖民了。
This is an odd question, because software engineers' salaries aren't that
high. I wrote a blog post on this, here: Software engineer salaries aren't
inflated-- at least, not for the 99%. Relatedly, I wrote a Quora answer on
the topic: see Michael O. Church's answer to Why do software engineers make
so much?
That said, I certainly have no illusions about the impossibility of an
industry-wide salary drop. It could happen. I think that, for the less
skilled programmers currently able to make $120,000 per year on the strength
of their degrees alone, it will. They're going to be hit harder than the
genuinely capable engineers, who might see a 20% drop at worst. Of course,
there's a duality to these sorts of things. Where there are pain and
obscurity, there's also opportunity. Google accelerated in the dead times of
2002-4 and Facebook was founded in 2004, when society at large had little
faith in startups. (I knew people who turned down 2004-era offers from
Facebook to take shitty, 100-hour-week, "analyst" positions at Goldman Sachs
. Startups were way out of style until about 2007.) Now, startups are "cool"
again, and a bunch of dogshit is getting funded and it's harder than ever
for genuine talent to be noticed. So there is bad in good times, just as
there is good in the bad ones. My point is that this possible (and not
certain) impending salary drop isn't necessarily a bad thing, if it hits the
underqualified the hardest and burns out the underqualified executives
first. I could survive a drop to "only" $120,000 per year; these
narcissistic poseurs, currently making a quarter million in their VP/NTWTFK
(Non-Technical Who-The-Fuck Knows) positions, couldn't.
If salaries crash, we're going to see a decline of interest in programming,
and here's what's going to happen. The less-capable programmers are probably
going to be flushed out of the system. The capable ones like me will be
hurting, but it'll be a small salary dip: probably more like stagnation than
a fall. Socially astute people will realize their improved leverage with
the thinning crowd and grab more opportunities. When no one's making money,
you want to optimize for interesting work and for status; when the market
turns favorable again, you have better odds of having the "overnight success
that takes 8 years" and hitting it big. If you want to gain a couple levels
in "data science" or management, or to cross the chasm and become a VC, now
's not the best time to ask, because the field's a bit congested with people
willing to accept maltreatment because they expect to be dot-com
millionaires or startup CEOs inside of four years. If the market crashes
tomorrow, then two years from now will be a great time (if you're good) to
start pushing forward on the autonomy/status axis. It'll be hard as hell to
get a raise, but easier to get the title of "Chief Architect", or to get
introductions to VCs (who won't be considered "cool" or inaccessible, post-
crash). Six years later, when times are good again, you can win.
I don't see "new people entering the field" as a major issue, because if you
give programmers autonomy and resources they tend to advance the state of
society, and that creates jobs. It's not about the number of people, but the
conditions they're willing to accept. In fact, what keeps our salaries
mediocre and our status and autonomy so low is that we're terrible at
organizing around our interests. We should have a union or professional
guild, and we don't. We should have private talent agents, and be organized
enough that we can afford to pay them. We should have an employment field
where binding mandatory arbitration, fast firing without severance,
aggressive non-competes and "we own your side projects" clauses don't exist,
and where we have legal representation at employer expense in the event
that our rights are violated. If a history major with stronger social skills
, who discovered a passion for technology in her mid-30s, comes in to our
fold, I think that's great. We need more diversity of thought and
perspective. We need more socially skilled people who identify as
technologists and engineers so we can lead from within instead of being led
by the clueless, short-sighted McKinsey culture that is currently colonizing
us (see: The Damaso Effect).
Now, the H1-B program has depressed wages, and that has a lot of people
scared. The issue isn't classical wage depression, though. Once you bring
someone to the Valley, his wage expectations are the same as anyone else's
after 5 years. Someone who's ambitious enough to come to this country isn't
going to be happy to be mistreated and underpaid just because he's from a
poor country. The major failing of the H1-B program is that it creates a
pool of essentially captive labor, because losing your job means you'll
probably have to leave the country, and quickly, and international moves
without a relocation or repatriation package (and startups don't pay for
those, because most of them are shitty) are terrifying. The H1-B program was
designed to bring in high-talent outliers, and instead, it's being abused
to create a pool of slave labor. Its effect on wages at the lower end is
small (H1-B workers still get middle-class salaries; the program isn't
nearly as abusive as, say, Japan's programs for Southeast Asian tech workers
) but the program's general effect on programmer autonomy and status is
devastating (and that hurts our salaries, later on). I have no problem with
the number of people brought in by this program (in fact, there aren't many
H1-B slots) but, if we're going to have that, then admission to this country
has to be unconditional with regard to employment; once you have it, you're
good for a certain number of years no matter how often you change jobs.
Fixing the H1-B program should be a high priority. And that's, again,
something that probably won't happen until we (as technologists) bring a
more diverse set of people into our fold.
Okay, with all that said, time to answer the question: I don't know. Knowing
that there is a bubble is easy (we're in one) but separating the genuine
changes from the bubble effects is very hard. (See: the 1990s bubble. The
Internet really was a BFD.) Predicting when a bubble will bust and how is
probably impossible. Software engineer salaries aren't that high, so they
could credibly increase or decrease in the next five years. I wouldn't worry
about it, because I don't see them changing by enough to sway career
decisions-- at least, not if you're capable and passionate. Whether the
economic tide goes for or against our industry, we need to get much better
at organizing around our interests. If times go well and there's more
prosperity, we deserve to share in it, instead of letting well-connected,
pedigreed McKinsey assholes take everything. If they go poorly, we want to
be protected. Either way, we need to get better-organized, and there's no
time to waste.
