A couple of days ago, someone on Facebook asked me what I thought of this post by Paul Graham. My first thought was, “Who is Paul Graham, and why should I have an opinion on the piece?” Part of the answer presented itself as I read his post. Graham is one of the founders and the head of Y Combinator, the largest seed money firm for tech company start-ups, and his post was answering some unspecified charges about sexism. That’s worth paying attention to.
So I did that. Moreover, I did that without looking for the original charges of sexism. The vast majority of this was written without that knowledge, and once I did know what the to do was about, it added to my analysis, but it didn’t invalidate anything I’d already written. I’d recommend saving that link for last as you read as well.
The Good News About Paul Graham
Graham’s post says he wants to be part of the solution, and it generally backs that up. He doesn’t treat the gender disparity as a function of some baseline difference between men and women. He acknowledges that bias exists in the industry. His company’s numbers suggest he is helping to an extent, though I think he’s using the wrong benchmark. (More on that in a bit.) The post contains a strong statement that people who don’t do what he does to be inclusive of female founders are losing out monetarily.
The Bad News About Paul Graham
This is largely bad news about the tech sector in general, but Graham’s arguments both in the original interview and here are about 10 years out of date in important ways. I say 10 years because “get girls interested” and female role model campaigns have already been going on in STEM for long enough that we’ve been measuring the results from them for several years. We’ve known for a while that the answer isn’t that simple.
The Good News About Paul Graham’s Ideas
Programs like those Graham has suggested do create interest in kids. Exposure to computers and to programming increases interest and confidence. These programs do turn out girls who want to program for a living. They increase participation in these subjects in grade-school classes, particularly in elementary and early middle school. Having role models does continue to help beyond that point, though those role models may have to be closer than a tech start-up to make the difference he wants. He is also right that having more female founders works toward normalizing the idea of female founders.
The Bad News About Paul Graham’s Ideas
These programs don’t do nearly as much as expected to increase the numbers of women in STEM careers and very little to increase the number of women in the top ranks of STEM careers. At every stage further from those “get girls excited about science” programs, the disparity between men and women grows. And when you work with the women in the field, you find out the difference (men drop out at every stage too, just not as many) is largely due to biased gatekeepers, sexual harassment, programs/industries that are tailored for people who never get burdened with caretaking activities, and the general, vague air that they don’t belong there. While Graham gives these a passing mention, his focus is elsewhere.
The Good News for Paul Graham’s Company
The women he’s investing in are very likely to outperform the men he’s investing in. Because YC has funded women overall at about the same proportion as they exist in the programming pool, their female founders are skewed (at 11% instead of 50%, heavily so) toward the top ranks of skill and dedication. This happens because the female programming pool has already been highly selected for both those things. If you’re not dedicated, you don’t put up with the crap. If you’re not very good, you don’t get past the gatekeepers.
The Bad News for Paul Graham’s Company
He’s missing some of the peak talent. Because he’s funding female founders only at rates roughly proportional to a highly biased field, he’s almost certainly funding men who will not succeed as well as female candidates he’s not funding. This is all the more true because other seed companies are not even funding proportionally, leaving a bigger pool of top women. And it will be exacerbated by his stated–gendered–biases against, say, start-ups run by women with small children and people who have coded for years as a goal in itself rather than as a means to an end. Working his way, he’ll miss people with the passion and insight to solve problems for untapped markets. All this may self-correct somewhat over time, since it looks like their female funders percentages are increasing annually. If he keeps examining his irrational biases, he should get there all the sooner.
So basically, yes, he’s still part of the problem, though not nearly as big a part of the problem as many. It’s costing him as well as women in the industry, even though he’s doing it out of ignorance rather than malice. However, if he’s going to pontificate about women in tech in public, I’d argue he has a duty to his own ideals, if nothing else, to get educated on the subject.
For more reading on the topic that’s accessible online, check out Heather Dryburgh’s “Underrepresentation of Girls and Women in Computer Science: Classification of 1990s Research” (pdf) and Kenneth Koput and Barbara Gutek’s Gender Stratification in the IT Industry: Sex, Status and Social Capital.