On the mess at Google
Aug. 8th, 2017 10:52 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
drwexI was startled to read (on one of my tech mail lists) that a person felt firing the Google employee who circulated the position piece on the inferiority of women (henceforth G) was "petty." It was not petty, nor done without thought.
I've never worked at Google, nor do I have any insight into their internal processes. But I've worked at a lot of tech companies and talked to many people at many others; I believe my experience generalizes. I have also worked at companies that have made the transition from private to public, which is significant. And finally, I have a background in cognitive science and experience working in Compliance. Both are relevant here.
My first response to hearing about this screed was "Oh no, not again." This is the same wrong pseudo-science BS that got Larry Summers in trouble at Harvard a decade ago [1]. It is true that people born bio-female have brains that are structurally different from people born bio-male. It is also true that people who engage in daily meditation and other mindfulness practices have brains that are structured differently from people who do not engage in these practices. Anyone who tells you they can draw a straight line from these observable differences to anything about the general population is blowing smoke (to put it politely).
The truth of the matter is that we still don't understand how brain structures translate to observable external behaviors across a population. We don't understand why some football players develop visible dementia and other symptoms - and others do not - while nearly 100% of post-mortem brain analyses of football players show significant CTE. We don't understand how early exposure to trauma, or deficiencies in juvenile diet affect behaviors in adulthood. We know that reading to children when they are young seems to have benefits that persist across decades of life and those benefits seem to transcend gender and socio-economic boundaries. Put _that_ in your brain structure debate and explain it. (Spoiler: we can't, yet.)
So on the science, G was just flat-out wrong. The fact that the manifesto was written in pseudo-scientific terms was a ploy to make you think G had done his homework, and played on peoples' confirmation biases. If you were inclined to agree with the conclusion then the structure made it seem more plausible. But it's not - it's bullshit.
If spouting bullshit was a fireable offense, tech companies would be vast wastelands. Everyone from the most well-known CEOs down to the just-out-of-college programmer spouts bullshit from time to time. Sometimes people even spout bullshit on internal mailing lists or chat boards. Good tech companies have policies about these things, and train their employees on those policies. Regardless, every tech company makes you sign an agreement to follow company policies and there are consequences for violation of policies.
Tech companies tend to be looser and less structured than companies of comparable size in other industries, but you still have policies and they still ought to be enforced. [2] So that means you can't just spout arbitrary bullshit on company time and with company resources. Most tech companies permit employees to use their equipment (computers, networks, printers, etc.) with agreement or understanding that such use will be in moderation and not interfere with the operation of the business. You can print out a copy of your vacation itinerary, but printing a copy of the entire backpacker's guide to the Appalachian trail will likely get you in trouble.
Companies are not free-speech zones. They can, and do, ban speech on particular topics (e.g. religion or abortion) and discriminate against certain viewpoints (e.g. criticizing the company). You agree to these restrictions as a condition of your employment there. Spouting bullshit might or might not cross into these forbidden areas - it likely depends on the particular sort of bullshit you spout. Companies probably don't care if you loudly proclaim that Michael Jordan was better than Larry Bird in every way, so long as you're not stopping people getting work done while doing it. But they may well care if you're encouraging fellow employees to spend money on fantasy basketball leagues.
Google CEO statements after the firing seem to be trying to hedge this, with Pichai talking about freedom of expression. Google may indeed encourage discussion of issues such as the effectiveness of diversity programs. But that's almost entirely beside the point - the difference in statements between "I don't think my company's diversity program encourages a wide enough range of viewpoints" and "I think a large chunk of my coworkers are biologically unsuited to the jobs they hold and only got those jobs through misplaced policies" is exactly the non-free speech distinction. In the wider United States, both are statements of opinion and protected speech. In a corporate environment, one expresses an unacceptable viewpoint and is prohibited.
ETA: I've read and reread the Pichai letter and I think I didn't give him enough credit on first writing. He's treading a careful line and I think he does a pretty good job of it. A lot of people (dare I say brogrammers?) are going to spend their time whining about "free speech" or similar regardless. But I think the main point stands: companies are not free-speech zones, and Google's policy is to encourage some kinds of expression, not any kind.
The vast majority of tech companies are private; a private company gets to do a lot of things with very few externally imposed rules. Once a company goes public, though, a whole host of rules start to apply. Public companies have to obey laws that private companies do not. They have to make reports that private companies do not. And they have to have policies and procedures that private companies do not.
For example, I was once fired from a job at a private company directly by the CEO. This person decided his company no longer needed UI Design and he fired everyone with that job. In my case that was over the objections of my boss, and my grandboss. But they didn't get a say in it - the CEO ran the company as his private fiefdom. A public company can't do that - there needs to be a procedure for firing an individual and even though most tech employees are "at will" - meaning the company doesn't have to have just cause nor give you warning - being a public corporation imposes some restraints on that.
A good public company will have (often as part of its HR, though in some industries such as banking and finance it's separate) Compliance people. These are the folk whose job it is to make sure that all the rules get followed. All the i's dotted and t's crossed. Because being sued for "wrongful termination" is still a very big and very expensive thing. The bigger and more public a company is, the more likely they are to be subject to such suits and it's Compliance's job to put the company in a good position to defend against those suits if need be. More importantly, it's their job to advise the company on how to avoid ever being in a position to be sued. [3]
One of the lesser-known things about public companies is that the law regulates who "speaks for" the company. In a private company the person in charge can say "I don't care what XYZ said, this is my company and I speak for it." And that sticks. In a public company, anyone in a "position of responsibility" may be assumed to speak for the company. Courts have ruled that anyone who has managerial authority may be assumed to be in a position of responsibility and therefore their statements are the company's positions. This is one reason why people ladle their statements with disclaimers (often written by the company's lawyers) and why companies control things like who gets to go speak at conferences, who talks to the press, and so on.
G got fired, I would bet, not for spouting bullshit. G got fired for violating company policies against creating a hostile work environment, and likely for damaging the reputation of the company. Although the bullshit manifesto was posted to internal lists it got out and was widely circulated outside. At a minimum, people had to ask the question "Is this Google's position?"
From the moment I heard about this there was no question in my mind that G was getting fired. However, because Google is a public company and because it's highly exposed they weren't simply going to frog-march him to the door, as Yonatan Zunger suggested in an otherwise well-reasoned public response. My bet is that G got a termination agreement, which likely included some amount of money in exchange for which he agrees (a) not to talk about this settlement, and (b) not to dispute his firing on pain of losing this cash. This is what Compliance does for you - it tells you how much it would cost if G did file a wrongful termination lawsuit and helps you figure out how much you should offer him in exchange for a promise not to do that. It took a while to draw up that termination agreement and to get the right managers in a room to be briefed on it. Since G's name is public, it won't be long before reporters come knocking on his manager's door, and his manager's manager's door looking for juicy bits. That has to be clamped down, completely, and everyone inside Google in that management chain has to be on the same page about this. That takes time.
Google is already on the losing side; if nothing else, they lose all the investment and training they put into G and they have to incur the expense of finding a replacement. There is reputational damage and diminishment of talent in their hiring pool. They will likely lose some current valuable employees who see that other people inside Google supported G and don't want to work in that kind of environment - see above about hostile workplace. The cash they'll pay to G is small potatoes compared to their already considerable losses.
If I was in charge of this at Google I would be terminating my agreement with their PR firm. Their public silence on this has done them a world of hurt and their PR firm should not have let that happen. [4] As I noted, far too many people are now asking questions like: "Is this Google's policy? Is this what they actually think? Are their diversity and inclusiveness initiatives nothing more than window dressing?"
I don't think it matters what you believe the answer to those questions to be - what matters is that those questions ought never to be asked. The PR people should be putting out very clear statements (or writing such statements for the management to deliver) that make it 100% crystal clear Google stands behind its initiatives and all its employees.[5]
[1] Note this is not why Summers was fired; see Cathy O'Neill's explainer from a few years back about what actually got him fired. But that's a digression. Summers was wrong, and G is wrong.
[2] There's a whole lot to say about what's wrong with tech companies in these areas but that's sort of aside. Take as given please that I think places like Uber are a cancer and should be cleansed with fire if nothing other than serving as a warning to everyone else.
[3] My actual background is in financial compliance - NASD, SEC and so on - but the principles are the same across industries. Financial compliance is just more complicated and more expensive.
[4] It's possible that the PR firm did advise them on this and Google didn't take that advice - I have no inside insight.
[5] Google's stock price doesn't per se exist since it's part of Alphabet, Inc and listed on the NASDAQ that way. In the last five days the price has been pretty stable so maybe this is a tempest in the tech teapot and the rest of the world doesn't give a hoot. If the stock was nosediving you'd be hearing a very different tune from Pichai.
I've never worked at Google, nor do I have any insight into their internal processes. But I've worked at a lot of tech companies and talked to many people at many others; I believe my experience generalizes. I have also worked at companies that have made the transition from private to public, which is significant. And finally, I have a background in cognitive science and experience working in Compliance. Both are relevant here.
My first response to hearing about this screed was "Oh no, not again." This is the same wrong pseudo-science BS that got Larry Summers in trouble at Harvard a decade ago [1]. It is true that people born bio-female have brains that are structurally different from people born bio-male. It is also true that people who engage in daily meditation and other mindfulness practices have brains that are structured differently from people who do not engage in these practices. Anyone who tells you they can draw a straight line from these observable differences to anything about the general population is blowing smoke (to put it politely).
The truth of the matter is that we still don't understand how brain structures translate to observable external behaviors across a population. We don't understand why some football players develop visible dementia and other symptoms - and others do not - while nearly 100% of post-mortem brain analyses of football players show significant CTE. We don't understand how early exposure to trauma, or deficiencies in juvenile diet affect behaviors in adulthood. We know that reading to children when they are young seems to have benefits that persist across decades of life and those benefits seem to transcend gender and socio-economic boundaries. Put _that_ in your brain structure debate and explain it. (Spoiler: we can't, yet.)
So on the science, G was just flat-out wrong. The fact that the manifesto was written in pseudo-scientific terms was a ploy to make you think G had done his homework, and played on peoples' confirmation biases. If you were inclined to agree with the conclusion then the structure made it seem more plausible. But it's not - it's bullshit.
If spouting bullshit was a fireable offense, tech companies would be vast wastelands. Everyone from the most well-known CEOs down to the just-out-of-college programmer spouts bullshit from time to time. Sometimes people even spout bullshit on internal mailing lists or chat boards. Good tech companies have policies about these things, and train their employees on those policies. Regardless, every tech company makes you sign an agreement to follow company policies and there are consequences for violation of policies.
Tech companies tend to be looser and less structured than companies of comparable size in other industries, but you still have policies and they still ought to be enforced. [2] So that means you can't just spout arbitrary bullshit on company time and with company resources. Most tech companies permit employees to use their equipment (computers, networks, printers, etc.) with agreement or understanding that such use will be in moderation and not interfere with the operation of the business. You can print out a copy of your vacation itinerary, but printing a copy of the entire backpacker's guide to the Appalachian trail will likely get you in trouble.
Companies are not free-speech zones. They can, and do, ban speech on particular topics (e.g. religion or abortion) and discriminate against certain viewpoints (e.g. criticizing the company). You agree to these restrictions as a condition of your employment there. Spouting bullshit might or might not cross into these forbidden areas - it likely depends on the particular sort of bullshit you spout. Companies probably don't care if you loudly proclaim that Michael Jordan was better than Larry Bird in every way, so long as you're not stopping people getting work done while doing it. But they may well care if you're encouraging fellow employees to spend money on fantasy basketball leagues.
Google CEO statements after the firing seem to be trying to hedge this, with Pichai talking about freedom of expression. Google may indeed encourage discussion of issues such as the effectiveness of diversity programs. But that's almost entirely beside the point - the difference in statements between "I don't think my company's diversity program encourages a wide enough range of viewpoints" and "I think a large chunk of my coworkers are biologically unsuited to the jobs they hold and only got those jobs through misplaced policies" is exactly the non-free speech distinction. In the wider United States, both are statements of opinion and protected speech. In a corporate environment, one expresses an unacceptable viewpoint and is prohibited.
ETA: I've read and reread the Pichai letter and I think I didn't give him enough credit on first writing. He's treading a careful line and I think he does a pretty good job of it. A lot of people (dare I say brogrammers?) are going to spend their time whining about "free speech" or similar regardless. But I think the main point stands: companies are not free-speech zones, and Google's policy is to encourage some kinds of expression, not any kind.
The vast majority of tech companies are private; a private company gets to do a lot of things with very few externally imposed rules. Once a company goes public, though, a whole host of rules start to apply. Public companies have to obey laws that private companies do not. They have to make reports that private companies do not. And they have to have policies and procedures that private companies do not.
For example, I was once fired from a job at a private company directly by the CEO. This person decided his company no longer needed UI Design and he fired everyone with that job. In my case that was over the objections of my boss, and my grandboss. But they didn't get a say in it - the CEO ran the company as his private fiefdom. A public company can't do that - there needs to be a procedure for firing an individual and even though most tech employees are "at will" - meaning the company doesn't have to have just cause nor give you warning - being a public corporation imposes some restraints on that.
A good public company will have (often as part of its HR, though in some industries such as banking and finance it's separate) Compliance people. These are the folk whose job it is to make sure that all the rules get followed. All the i's dotted and t's crossed. Because being sued for "wrongful termination" is still a very big and very expensive thing. The bigger and more public a company is, the more likely they are to be subject to such suits and it's Compliance's job to put the company in a good position to defend against those suits if need be. More importantly, it's their job to advise the company on how to avoid ever being in a position to be sued. [3]
One of the lesser-known things about public companies is that the law regulates who "speaks for" the company. In a private company the person in charge can say "I don't care what XYZ said, this is my company and I speak for it." And that sticks. In a public company, anyone in a "position of responsibility" may be assumed to speak for the company. Courts have ruled that anyone who has managerial authority may be assumed to be in a position of responsibility and therefore their statements are the company's positions. This is one reason why people ladle their statements with disclaimers (often written by the company's lawyers) and why companies control things like who gets to go speak at conferences, who talks to the press, and so on.
G got fired, I would bet, not for spouting bullshit. G got fired for violating company policies against creating a hostile work environment, and likely for damaging the reputation of the company. Although the bullshit manifesto was posted to internal lists it got out and was widely circulated outside. At a minimum, people had to ask the question "Is this Google's position?"
From the moment I heard about this there was no question in my mind that G was getting fired. However, because Google is a public company and because it's highly exposed they weren't simply going to frog-march him to the door, as Yonatan Zunger suggested in an otherwise well-reasoned public response. My bet is that G got a termination agreement, which likely included some amount of money in exchange for which he agrees (a) not to talk about this settlement, and (b) not to dispute his firing on pain of losing this cash. This is what Compliance does for you - it tells you how much it would cost if G did file a wrongful termination lawsuit and helps you figure out how much you should offer him in exchange for a promise not to do that. It took a while to draw up that termination agreement and to get the right managers in a room to be briefed on it. Since G's name is public, it won't be long before reporters come knocking on his manager's door, and his manager's manager's door looking for juicy bits. That has to be clamped down, completely, and everyone inside Google in that management chain has to be on the same page about this. That takes time.
Google is already on the losing side; if nothing else, they lose all the investment and training they put into G and they have to incur the expense of finding a replacement. There is reputational damage and diminishment of talent in their hiring pool. They will likely lose some current valuable employees who see that other people inside Google supported G and don't want to work in that kind of environment - see above about hostile workplace. The cash they'll pay to G is small potatoes compared to their already considerable losses.
If I was in charge of this at Google I would be terminating my agreement with their PR firm. Their public silence on this has done them a world of hurt and their PR firm should not have let that happen. [4] As I noted, far too many people are now asking questions like: "Is this Google's policy? Is this what they actually think? Are their diversity and inclusiveness initiatives nothing more than window dressing?"
I don't think it matters what you believe the answer to those questions to be - what matters is that those questions ought never to be asked. The PR people should be putting out very clear statements (or writing such statements for the management to deliver) that make it 100% crystal clear Google stands behind its initiatives and all its employees.[5]
[1] Note this is not why Summers was fired; see Cathy O'Neill's explainer from a few years back about what actually got him fired. But that's a digression. Summers was wrong, and G is wrong.
[2] There's a whole lot to say about what's wrong with tech companies in these areas but that's sort of aside. Take as given please that I think places like Uber are a cancer and should be cleansed with fire if nothing other than serving as a warning to everyone else.
[3] My actual background is in financial compliance - NASD, SEC and so on - but the principles are the same across industries. Financial compliance is just more complicated and more expensive.
[4] It's possible that the PR firm did advise them on this and Google didn't take that advice - I have no inside insight.
[5] Google's stock price doesn't per se exist since it's part of Alphabet, Inc and listed on the NASDAQ that way. In the last five days the price has been pretty stable so maybe this is a tempest in the tech teapot and the rest of the world doesn't give a hoot. If the stock was nosediving you'd be hearing a very different tune from Pichai.
no subject
Date: 2017-08-08 07:06 pm (UTC)So thanks -- it's useful to understand these private / public differences. Especially since I first found out about the brouhaha from Zunger's article, and had taken that detail at face value...
no subject
Date: 2017-08-08 08:25 pm (UTC)Almost all of these fall under "clear and present danger" policies and the company will document how the firee presented such a danger; if they're smart they've been building a case for some time and didn't get caught by surprise.
Next time we're face to face ask me about the Big Bank and Paris Hilton...
no subject
Date: 2017-08-09 02:06 am (UTC)Whoever has been leaking internal blog posts, etc, directly to Vox Day, if caught, will be gone instantaneously.
no subject
Date: 2017-08-09 03:05 am (UTC)no subject
Date: 2017-08-09 02:05 am (UTC)no subject
Date: 2017-08-09 03:07 am (UTC)Sorry about the life-being-made-hell thing.
no subject
Date: 2017-08-09 04:34 pm (UTC)There is also a lot we don't know about how muscles work, and a lot we don't know about how the detailed structure of our muscles translates into the ability to lift weights. But I think we can still say with confidence that on average, men are stronger than women.
We can also say (with a level of confidence that is somewhat lower, but still extremely high) that on average, men are stronger than women for reasons that have something to do with genetics. You might say that there's no way empirically to distinguish that hypothesis from the hypothesis that boys are encouraged, moreso than girls, to build up their strength. And here I would say: yes, you're right---that's why I said we can make the genetic connection with "somewhat less confidence" But I think the level of confidence, while "somewhat less", is still extremely high.
It is extremely high for a great many reasons, one of which is that it makes sense, given what we know of the environment in which we evolved, for there to have been selection pressures for men to be stronger than women.
Likewise: The observed variance of IQ is higher in men than it is in women. I believe the evidence is quite clear on this. There's a lot we don't know about the relation between brain structure and IQ, but that need not prevent us from making use of the things we *do* know. One of the many relevant things we know is that in a world where one sex invests a lot more than the other in the process of gestation, there are selection pressures for that sex to have less variance in a great many traits. That is a relevant fact, and just because we don't have all the facts about brain structure that we'd like to have, it does not follow that we should ignore known relevant facts.
I have not read G's document, so this is not a defense of it. It is meant only as a defense of the proposition that it is possible to be pretty confident of some things even when we are pretty ignorant about a lot of related things, which is what I understand you to be denying.
no subject
Date: 2017-08-09 06:58 pm (UTC)No, what I said was that you cannot draw a line from any observable difference in brain structure and any observed (or unobserved) behavior in a population with that structure and say "See, the structural differences are the reason for the behavioral differences." This is the core of the "biological bases" argument which takes the typical form "women are unsuitable for X because of different brain structure."
I think we can still say with confidence that on average, men are stronger than women.
I think that is a sloppy sentence. You can say with confidence that "people with more muscle mass are on average stronger than people with less muscle mass." You can likewise point out that people who have average or higher testosterone and androgen levels will develop more muscle mass. Now we're describing a causative chain; this is precisely what's missing in the discussion of brain structure. My point is not merely that there's a lot we don't know about brains - my point is that among the things we don't know is anything about how observable differences in brain structure relate to generalizable differences in population behavior.
it makes sense, given what we know of the environment in which we evolved, for there to have been selection pressures for men to be stronger than women
I am always very suspicious of the "evolution explains things" line of reasoning. We don't have first-hand evidence of what the evolutionary pressures were in the past; all we can see is who the evolutionary survivors were. Evolutionary explanations also fail to account for the persistence of prima facie counter-evolutionary traits. And finally, our notions of what actually happened are still being massively revised. It's only within the last few decades that we learned that genetic material moves cross-species in wildly unexpected ways, such as humans exchanging genetic material with our gut bacteria. It's only within the last decade that we came to understand that homo sapiens in fact interbred extensively with other coextant hominid species, and only within the last two years that we found really good evidence for the idea that homo sapiens evolved in many locations rather than in one.
Most clearly, though, evolutionary explanations depend on the idea that evolution behaves in what we think of as a current, rational manner, given the set of cues we take to be important. As I understand it, things aren't that simple, and evolution often has unexpected consequences.
The observed variance of IQ is higher in men than it is in women. I believe the evidence is quite clear on this
Did you mean variance in the sense that men display a wider range of scores on this test than women do? Are you confident that you have a large enough sample size to say anything reliable about people who score more than a standard deviation or three away from the mean? Are you confident this test repeatably measures anything meaningful? I'm not confident of either of those things.
in a world where one sex invests a lot more than the other in the process of gestation, there are selection pressures for that sex to have less variance in a great many traits
I'm not a biologist so I'm not in a position to say whether that's so or not, but I'm curious what its relevance is. I understand you to be saying that "because reproduction, child-bearing members of a species will have less variance from the mean on a great many traits."
What's the "and therefore..." from this?
no subject
Date: 2017-08-09 09:05 pm (UTC)no subject
Date: 2017-08-10 03:55 pm (UTC)no subject
Date: 2017-08-10 05:16 pm (UTC)Are you able to parse that statement? Are you able to agree with it, beyond just playing along?
no subject
Date: 2017-08-10 06:03 pm (UTC)(I'm still playing along because I reject a gender binary but I can easily imagine sampling from people who fit what I think you're pointing toward.)
no subject
Date: 2017-08-10 10:01 pm (UTC)"Able to parse" was not meant as an insult; I understood you to be saying that you were unable to parse a statement in my original comment that seemed to me to be structurally identical to this one. I had no idea what problem you were having so I had no idea whether I'd overcome the problem. Therefore I asked.
So:
1) We agree, I think, that one can reasonably believe the statement that "men are on average stronger than women", even though we (or at least I, and am guessing you) know very little about what happens at a cellular level to make some people stronger than others.
2) Therefore knowledge of what happens at a cellular level is not a requirement for reasonably believing a statistical assertion about gender differences.
3) Therefore, the fact that neither you nor I (nor perhaps anyone) knows very much about the details of how brain structure influences skills and behavior ought not, by itself, prevent us from believing statistical assertions about gender differences, even when those differences arise from processes we do not understand.
4) I understood your assertion to be that because we do not understand much about how brain structure affects gender differences, we can not reasonably believe that the variance of IQ is higher in men than in women.
5) Give 3), and given my understanding in 4), your entire argument was invalid. In fact, glaringly invalid. In fact, so glaringly invalid that I'm kind of shocked that I have to explain this.
6) Having established that we *can* reasonably believe that IQ variance is higher for men than for women, the question is: *Should* we believe it? The answer to that seems to me to be clearly yes, for the same two broad classes of reasons that we believe that average strength is higher in men than in women. First, there's considerable empirical evidence. (My familiarity with this evidence is outdated, but it was quite considerable last I checked, and I have no reason to think the weight of the evidence has changed much.) Second, we have --- and this is not irrelevant --- some very rough idea of a mechanism that could have plausibly led to this difference.
7) You might say that a very rough idea of a mechanism is not much to go on, but it's a good part of what we go on when it comes to strength. Part of the reason I believe that men are stronger than women on average is that there is considerable empirical evidence, but another reason I believe it is that it makes evolutionary sense. I can't fill in the details of that evolutionary argument; I can't be sure it's even close to right, but the fact that it *could* be right gives me a little more confidence in the generalization. That is a reasonable thing when it comes to strength averages and equally reasonable when it comes to IQ variances.
8) Bottom line: I think it is extremely clear that the variance of IQ differs across the sexes. My understanding of your position is that not only do you dispute this, you dispute also that it's even reasonable to believe such a thing. But you believe analogous things when it comes to strength, and you believe them for analogous reasons. So you are being, at the very least, inconsistent, and possibly extremely dense.