The reality is that the stuff we're talking about here would, for the most part, be a firing offense at most companies.
For people who pride themselves on being analytical and meritocratic, it can be a difficult thing to realize that no amount of detailed analysis is going to help because you've been living in a bubble for most of your life -- you have only a small amount real data to work with. It can be even more startling when your meritocracy not ends up not only not really being one, but possibly impossible to achieve given the constraints of human nature.
and here we're talking about how perpetrating of stereotypes is a bad thing
edit: found a word for it in English - hypocrisy.
I'd like to just add a bit (edit: okay, a rather long-ish bit...) about how a reasonable, non-sexist guy could end up assuming that things are not that bad in the industry, whereas a reasonable women can have experienced more sexism than in other industries, and both can be "right". All without the typical (and IMO, rather cynical) assumption that guys are just playing along, encouraging it, or putting on blinders. And also (more importantly) without the assumption that men in the industry are any better or worse than in any other.
This should make both sides pause a bit before they scream about how unreasonable people on the other side are being...
For a lot of men, the skepticism is not over whether actions are sexist or not, but over how often they actually happen.
My argument is that this difference in perception is almost exclusively due to the extremely lopsided sex ratio in tech, not due to people in tech being any better or worse: I don't know exact numbers, but let's say somewhere around 10% of tech workers are female (in my experience it's even worse than that, but I don't know for sure, industry-wide).
Going with that number, that means that out of a random sample of interactions between other people that you (let's assume "you" are a non-sexist guy that can accurately recognize sexism when you see it) personally witness over the course of a career, only 18% will be between a man and a woman. As for the percentage of sexist guys (I'm making the simplifying assumption that a guy is either sexist or not-sexist - I could easily remove this restriction and replace it with a probability distribution, but it would needlessly complicate things)...I'm not sure about that, but let's aim high and say it's 20% (I don't think more than 20% of us would, for instance, show pantless pictures of ourselves, send harassing e-mails, proposition an intern, ask about kids at an interview, etc.). Even when a sexist guy interacts with a woman, we should probably assume that it's a reasonably small percentage of those interactions, maybe 10%, where he'd actually say or do something offensive, especially with someone else present.
[Again, all these numbers are pure fiction, placeholders for the purpose of demonstrating the extreme effect that the sex-ratio has, rather than figuring out anything in detail]
Put that all together, and let's say that you, a not-sexist guy, witness 1000 interactions between other people at work during some time period. By these estimates, only 3 of those interactions would be "sexist interactions." The problem seems rather small when you look at it that way, and in fact, it's small enough that statistical variation could mean that you never end up witnessing such interactions at all, even if they are happening at your place of work.
Now, the meat of the argument: consider, instead, the point of view of a woman. 18% of all of her interactions are with sexist men, and 1.8% of her interactions involve a guy acting sexist towards her. By the percentages, that's a sixfold increase over what you would notice as a man, even though the actual frequency of sexist behavior is the same.
I think that's where this sort of discussion breaks down: those of us that are not sexist, but are not women, see an apparent level of sexism that is six times lower than what women observe in their own work interactions, and that's arguably the difference between the perception that sexism is pervasive and oppressive, versus barely worth considering. And it's all due to the sex ratio - if it was 50/50, then the percent of interactions that are sexist that men observe (holding the other numbers the same) would be 50% * 20% * 10% = 1%, and what women observe in their own interactions would be 50% * 20% * 10% = 1%, so there would be an equal perception of sexism.
We all need to keep this in mind to bring some sanity these sorts of discussions: girls, it's not that the men that doubt this are assholes or privilege-denying-misogynists, it's that they truly, honestly witness sexism less often than they would if they worked in a field with a more balanced gender ratio, and it's not necessarily that they're complicit or ignoring it. And guys, the women aren't being whiny or weak, they truly, honestly experience a higher percentage of sexist interactions than they would in a field with more women.
FWIW, all of this applies to any minority situation, and is (IMO) one of the biggest unrealized factors that leads to disagreement whenever these arguments come up.
Another factor that contributes to male skepticism is that many of us have experience in companies that are extremely intolerant of sexism. Like Thomas said, most of what we're talking about would be firing offenses at most companies. I know if my wife ever witnessed this sort of thing at BigCo, the guy would've been fired on the spot.