Log in

No account? Create an account
29 December 2013 @ 05:16 pm
Interesting trade-offs in rights  
You've probably seen Paul Graham put his foot in his mouth about how it's hard to find brilliant woman technologists because they don't start programming early -- which, statistically, they tend not to.

Paul Graham is primarily making an observation -- he's a venture capitalist, and he doesn't find that people who didn't start hacking early are the kind of technologists that make him a lot of money. That doesn't specifically make his comments good or bad, excusable or inexcusable. But that's where he's coming from.

There's a huge outcry, of course, from people talking about how you don't need to start hacking at 13 to be an amazing programmer (and you don't.) The implication behind these things is that a woman with fewer years of experience should be every bit as good as a man with more years of experience.

Speaking as an old guy, I have no idea where to go with this. Because the argument that goes "more years of experience don't make you better!" makes me look like a pretty bad choice of employee. I'm a lot finickier about working conditions than when I was younger with fewer years of experience, for instance.

To be fair, startups already tend to discriminate against old guys (Paul Graham and Y Combinator less than many). And getting more young women in startups would be good all 'round.

But it looks an awful lot like this argument is a straight-up tradeoff of their advantage for older workers'.

I have quite a lot of advantage, and I can give some up. But "we should discriminate less against women in tech and more against old people" is still kinda mixed, you know?
IANALqueen_elvis on December 30th, 2013 03:17 am (UTC)
Are we at the point where 36 is old?
IANALqueen_elvis on December 30th, 2013 03:18 am (UTC)
Anyway, startup culture doesn't want you. It wants 22-year-olds who are willing to work 12-hour days. Being in a position to be more choosy is a good thing.
Noahangelbob on December 30th, 2013 03:23 am (UTC)
Looks like you answered yourself :-)

In startup culture, 37 is old.

Though honestly, at this point, 37 tends to be old for a programmer.
meganmh75 on December 30th, 2013 06:52 am (UTC)
Maybe you should move to Seattle. 37 is still not old even in the start-ups. Well, not in the ones I see, anyway.

I think you need to distinguish between experience and child's play. It might be true that girls are less likely to start hacking around on computers during childhood, but I don't know that women are trying to turn to computers as 26-year-olds and claiming that the first five years of industry experience don't matter.

I can you hear you exclaiming that the pre-college hacking isn't just child's play, and you'd be right in some ways. Among other things, the lack of easy familiarity with technology is something that prevents women from going into STEM disciplines in college. For another, it truly does add up to some experience. However, you'd be wrong in some ways, too, in that the five years spent playing and experimenting before entering college really don't substitute for five years of experience developing in industry. Nor does that auto-didactic exploration of computer science substitute for a more rigorous and organizing exploration that a college education does.

I think that someone who doesn't recognize the value of many years spent developing products from start to finish, working with economical realities, and deadlines, leading and working in engineering teams, and generally becoming an expert in a trade is awfully short sighted. Those hot-shot kids aren't going to grow into something truly great and well rounded without a mentor to lead them down the path.

(How you want to address the beginning issue where girls are less likely to immerse themselves in technology as kids is a different issue. Whether you can create an environment where women can feel comfortable working twelve hours a day with a group of men is another issue. Take it a step further and see if you can convert those women to long term industry participants even as they face child-bearing and you'll be a genius. I'd like to think there weren't inherent gender differences here, but, I don't know that this is true. Its a very difficult situation.)
Krissyrightkindofme on December 30th, 2013 01:11 pm (UTC)
Mostly what he is doing is teaching our kids to start building video games at 4. He doesn't care that they weren't born with a penis.

I don't know a different way to immerse girls in the culture other than to just do it.
Noahangelbob on December 30th, 2013 05:05 pm (UTC)
I'd tend to argue that younger experience tends to translate to "real" experience at a rate of around 10%-40%, depending. That is, ten years of before-industry experience (say from age 8 to 18) would translate into between one and four years of effective industry-type experience, depending a lot on what they do. It's just barely *possible* it might convert at 80%-100% if what they were doing *was* real industry experience or basically equivalent, but I've only once met that in the wild (Shawn Brenneman, who literally hired on to a tech company at 14). As you say, the early stuff is usually more playful.

I'd also agree that, in general, getting some college CS at some point helps a lot. In my experience, the start-at-13 crowd tend to also do that, though. If you skip college then sure, that can have bad consequences -- not always, but often.

And sure, developing products start-to-finish with a large team is its own skill. The lack of a career of that is often why startups, when they outgrow the team-of-up-to-eight stage, add project manager(s) and/or eng manager(s). That's often the skill that those folks are meant to import, often with varying results. You'll also, more rarely these days, see startups try to hire a few graybeards like myself, also get some of that same skill and perspective. It's necessary for a long-term sustainable company... But not necessarily as important for black swan farming, or "disruption" as the tech companies more frequently refer to it.

This is another place where age discrimination and sex discrimination tend to hit head-on. In general, VCs want startup founders who are either quite young (say 18-24) or significantly older (40+). When I say "want," of course I mean "fund." But this gap makes sense if you assume VCs want young childless people or people whose kids are old enough that the other partner can do nearly all the work raising them. Which, of course, VCs do.

For the younger camp, the 1-4 year head start in experience is actually a big thing. For people with zero to six years of post-college industry experience (18yo does a startup with no experience versus a 24yo found a company at 18), a "+1-4 year" bonus is a big deal.

The problem is different with 40+ founders -- that means if they were programming at 13, that was in something like 1987. It could be argued that women who started programming at 18 or 25 should be much more equivalent there -- a woman with fifteen solid years of industry experience won't be at much disadvantage compared to a guy with, say, 22 years of industry experience and 5-10 years of "play" experience. Some, but not nearly as much.

Unfortunately, I don't see a lot of this other camp of women either -- the ones with fifteen solid years of industry experience. They exist and I've worked with them, but the ones I know don't seem terribly inclined to go found companies and try to get them funded.

So from Paul Graham's point of view, that's not any easier a problem to fix. If it's going to take us awhile to get more girls who start coding at 13 or 16, it will take longer to get a similar percentage of them with 15+ years of industry experience. This is a problem that my industry is having in general now: we want a lot of highly senior people with huge amounts of experience, and it's hard to suddenly fix that problem. As a rule, it takes 15 years to get somebody with 15 years of experience, even if you can occasionally cheat by finding somebody with 5 or 10 already.

The current popular proposed solution ("I don't care if somebody picked up a hard skill at 13, I'm more impressed at what they did last year", to quote my Twitter feed) seems to be either that programming experience before college doesn't count, or that programming experience more generally doesn't count.
geekgirlwarsop on December 30th, 2013 09:27 pm (UTC)
The reason that you don't see a lot of women in who have 15 solid years of experience in a computing career (which is to say: me) is that women in computing careers drop out of their technical career at a much higher rate than women in other careers do. There is some correlation between dropping out of a technical career and having a family, but that correlation is insufficient to explain why women in computing careers drop out at a higher rate than women in other careers.

I gave a talk about this mid-career problem at a USENIX conference earlier this year. Twenty minutes before I gave my talk, one of my female colleagues (same level as me, a year or two older than I am) let me know that she was leaving the company so that she could pursue her PhD, and all I could think was, "another one bites the dust". At my employer, there are ~10k people in R&D. There are ~400 at my level (Staff) or higher (Senior Staff, Principal). There are about 20 women at my level, and exactly 2 who outrank me. If I look at my company's executives, there are three women: CMO, SVP in legal, and SVP of HR. So all three of our female executives are not in technical roles. This is especially frustrating given that my company was co-founded by a woman, she was our CEO for many years, and people here still refer to her by first name only.

I got my CS undergrad in 1998. Of the women who I've kept in touch with from my uni years, I'm the only one who's still in a technical role. Some are in non-technical roles in technical companies (say, they've moved to marketing), but most have completely moved away from anything to do with tech.
Noahangelbob on December 30th, 2013 11:38 pm (UTC)
Yeah. It's clear that the dropout rate is pretty substantial in my gender too, but I wouldn't be surprised if it were lower than the one for women.

I can think of a number of reasons that might be -- many of them applying to either gender -- but all I really have is speculation. I don't know of any good study of the phenomenon with a methodology I could call "not awful." Again, not specifically for women, but for both genders.
meganmh75 on January 2nd, 2014 06:40 am (UTC)
Women tend to drop out of many careers that have been old-boys-clubs. Women get converted to tenure at lower rates, women make partner in law-firms at lower rates, women leave tech at lower rates. Certainly, some of this is inflexibility for family development. Some of it, i think, it fatigue from years of dealing with gender bias day-in and day-out.
I think, when we are looking at getting more equal gender representation in STEM, we need to look seriously at career paths, not just career initiation.
rbus: squirrelrbus on December 30th, 2013 03:56 pm (UTC)
After working with men and women developing and using computer technology for the past 30 years I can tell you that men and women are different when it comes to developing and using technology. Brains are different. Different things are important. Observed problems are different. Solving problems is different.

In a nutshell:
-- Menlings = "Kill the mammoth!"
-- Femalians = "Is it just a mammoth?"

It's why I always want to work on a problem with reasonable women. And why women should always want reasonable guys. It just makes things better - if you can find reasonable women and men. *That's* the hard part.

As far as experience goes... A smart youngster can learn and apply knowledge (literally) hundreds of times faster than me. Experience makes me slow down with an eye towards saving effort and time. I love being the guy that says "We tried something like this ten years ago and it was difficult because of A, B, and C. What can we do this time to avoid those problems?"

There's -nothing- more rewarding than a youngling saying "I would've never thought of that" AND knowing that part of your experience is now theirs.

I don't try to compete with the smart young ones. But I can augment their skills to make them better at what their best at right now. It's important for me to understand that I'm not as smart as I used to be. It's important for them to understand they're not as wise as they will be.

"We grow wiser as we grow older is because it's too hard to get smarter."
- My Dad -
Krissyrightkindofme on December 30th, 2013 05:33 pm (UTC)
:) I'm glad you are in the world.
rbus: hugrbus on December 31st, 2013 02:55 pm (UTC)
Well, thanks!
(Deleted comment)
Noahangelbob on December 31st, 2013 06:23 am (UTC)
What he actually said was that he could always be *fooled* by someone who looks like Zuckerberg, and he pointed out that it was an exploit (indicating he was stopping), not a real and current criterion.

His determination that people who make him a lot of money are those that started hacking early is actually caused by the fact that he pretty exclusively funds people like that. He didn't test his hypothesis.

Interesting. Do you say that on the basis of anything specific he's said or done? I ask because "exactly like Mark Zuckerberg" isn't a great description of the founders I've seen him fund.

He also points out that that is *not* the only way he checks himself, though the Valleywag article only glances on that -- he grades himself by "misses", which almost certainly means "people we decided not to fund who then went on to succeed." I'm not saying it's not confirmation bias (it clearly is), but not quite as bad as the condition you mention.

The argument isn't about the number of years of experience but whether we should focus exclusively on young men who look like Zuckerberg

If current VCs are wrong in a systematic way, and there's a way to be right-er in a systematic way, then they should be leaving some scary amount of money on the table. For instance, assume that VCs systematically overrate young white males from prestigious universities (likely). Assume that applicants who lack at least one of those adjectives (young, white, male, from-prestigious-university) are overall a better pool -- less overrated.

In that case, VCs who systematically favor "minus-at-least-one" candidates over "exactly-all-of-those" candidates should make more money -- they're buying goods that are less overvalued.

The interesting thing is that VC is slavishly devoted to following a tiny "A list" of VCs (e.g. Foundation Capital, Sequoia, KPCB). If any of those companies were getting major wins by choosing "minus-at-least-one-adjective" candidates, you'd expect a small avalanche of other VCs to follow that approach. For instance, Paul Graham actually has a bimodal distribution on "young" and seems to make money at it, which is likely affecting other VCs' take on "must be young." Similarly, Indian founders are becoming part of "white" as far as VC investment goes (or correct the fundable set to "white, Asian or Indian", which also works), and this is clearly a correction based on empirical data.

They don't respond all that fast, but they clearly respond in many areas -- again, Graham more than most because he invests less per company, invests in more companies, and specifically founded Y Combinator to test some assertions about previously-excluded founders. YC happened because he believed VCs weren't willing to go young *enough*, and he's reset the accepted curve on that adjective. Most VCs followed because YC made a bunch of money.

None of that says that Paul Graham isn't wrong. The process I described starts with N-1 VCs being wrong and the one outlier making a bunch of extra money.

But if there's one thing VCs are reliably good at, it's watching the financial returns of other VCs. It would be very surprising if lots of VCs *stayed* wrong in the presence of other VCs trying a financially better strategy, particularly an easy-to-assess one like picking different founder demographics.

Also: whether it's "about" years of experience depends a lot on how we justify, doesn't it? If the common sentiment becomes "it doesn't matter how early you start coding," that says something about years of experience whether we originally intended it to or not.
Camille, Destroyer of Worlds: pic#1093692skamille on January 1st, 2014 07:01 pm (UTC)
Discrimination is a self-sustaining system. If you don't invest in women/minorities, you see fewer successful companies run by underrepresented individuals, you don't learn what to look for. You can't prove a counterfactual, and you can't say that because these are the companies they chose to invest in they are the best. You didn't invest in those people, so you have no idea what might have happened if you did. If the whole system discriminates against a particular group, that group will find it hard to succeed, and looking for "missed opportunities" is not enough to identify those who would have succeeded and change your metrics. VCs are not immune from the structural influences of discrimination just because they happen to have money in the game; they're not computer simulations, and they don't actually have all that many data points to work with.

I personally think the most striking thing about PG's bias shows in the full interview, where he states that things are different now:
It is changing a bit because it's no longer so critical to be a hacker. The nature of startups is changing. It used to be that all startups were mostly technology companies. Now you have things like the Gilt Groupe where they're really retailers, and that's what they have to be good at because the technology is more commoditized.

I'm sorry, but what? Airbnb is no more of a tech startup than Gilt is. They took literally an existing business model (vacation rentals), put an app + reviews on it, and created a company. Gilt completely changed a model for selling inventory through technology. This is a pretty absurd exclusion to make; oh, well, there are women involved but it's not actually tech. Dude, it's as tech as one of your great successes.
Noahangelbob on January 1st, 2014 07:13 pm (UTC)
Sure, VCs are far from perfect that way. But are you suggesting that previous programming experience just isn't a big deal when founding a new software company? If so, what is? I mean it as a serious question. I *have* met nonprogrammers who have founded various kinds of software-backed companies, but in my experience it's uncommon -- for starters, if you don't program then it's also hard to *hire* good programmers, you know? Or should they be looking more for people with a few years of programming experience and a few years of solid management experience?

If a given group finds it hard to succeed, there are (at least) two ways that can manifest: equally-good people from those groups can find it hard to succeed, or it can be hard for people from those groups to get equally good. In the first case, VC looks like it's structured in a way to take advantage of that -- there should be that pool of equally-good people who one or more VCs could scoop up and make a ton of money with. In the second case (hard to get equally good), it's not clear to me that the right answer is for VCs to invest more in that group. Would you argue that they should?

I agree - Gilt Groupe is certainly every bit as technical as AirBnb. It's kind of a weird example for him to use, honestly. I talked to some of the Gilt guys at a Mongo conference a few years back and they're certainly in a "need bleeding-edge tech" business, at least in several ways.

DropBox, YC's other scary-big success, *is* more of a traditional crazy-high-tech play. That is, the problem they're solving is just really hard. But I agree that that's not typical at this point.
Camille, Destroyer of Worldsskamille on January 1st, 2014 08:12 pm (UTC)
No, I'm not suggesting anything at all about programming experience, at all, the comment I replied to didn't even talk about programming experience. I'm not even talking about software-specific companies. VCs are funding companies that go far beyond "software" these days, as "software is eating the world".

I'm am saying that I don't believe that it is wise to make assumptions like this:
But if there's one thing VCs are reliably good at, it's watching the financial returns of other VCs. It would be very surprising if lots of VCs *stayed* wrong in the presence of other VCs trying a financially better strategy, particularly an easy-to-assess one like picking different founder demographics.

I think you really underestimate the structural inequality out there, and I think you overestimate VCs and the "logic" that they must use.

If a given group finds it hard to succeed, there are (at least) two ways that can manifest: equally-good people from those groups can find it hard to succeed, or it can be hard for people from those groups to get equally good. In the first case, VC looks like it's structured in a way to take advantage of that -- there should be that pool of equally-good people who one or more VCs could scoop up and make a ton of money with. In the second case (hard to get equally good), it's not clear to me that the right answer is for VCs to invest more in that group. Would you argue that they should?

I think that equally-good people from the group are underrepresented due to structural issues, and saying that VC looks like it's structured in a way to take advantage of that is naive, because VCs are humans and steeped in the same biases that we all are (if not moreso). I also think that there are fewer people from the groups that have the opportunity to get equally good, but that isn't really the point, because we aren't even talking about a representative percentage of women/minorities from the current tech population getting funded. And thirdly, I think it is incredibly difficult to even define what "good" is. There are legions of absolutely horrible tech company leaders out there that keep getting funding despite well-known instances of abuse and likely mental illness. The system is not self-healing, it perpetuates its own stereotypes, and it is not logical to assume that it is the way it is because the current system is in fact the best system, or even an optimal system.

Edited at 2014-01-01 08:13 pm (UTC)
Noahangelbob on January 1st, 2014 08:37 pm (UTC)
And thirdly, I think it is incredibly difficult to even define what "good" is

If we're talking VCs, "good" is defined as "founding a company that makes a very large exit" (or likely to do so). No individual is going to have odds of doing that above maybe 15%, even in extreme cases, but most people have effectively zero odds of it. Our old heuristics for that involved being really good at programming computers, and possibly things like management and business. The new heuristics are (always) up for debate.

As far as structural inequality... Sure. I agree. But it would still be very, very weird if that meant no VCs invest reasonably in women. For instance, there are a lot of structural problems with the stock market valuing long-term value stocks (as opposed to short-term growth and volatile stocks). This results in a small minority of people like Warren Buffett taking advantage to make giant baskets of money. The structural inequalities continue despite this small number of people doing well by it... And so a small number of people *keep* making giant baskets of money.

I believe that VC is structurally designed to avoid that problem, to the extent you can reasonably do that.

Assumptions like "equally-profitable black founders are much less likely to be funded" or "equally-profitable female founders are much less likely to be funded" would result in a small number of mini-Buffetts funding black and/or female founders and making a lot of money off it, because the best black and/or female founders would come to them first, giving them more of the best opportunities. If the rest of VC didn't follow suit, those VCs would just keep making good money off mostly-overlooked opportunities.

If you'd like an example of that, you could even look at YC itself. They believed that "VCs invest too much per company, too late in the process" and "founders can reasonably be much younger, to a minimum of 18 years old." And they made a bunch of money off it, and VC is scrambling to integrate those beliefs now.

I don't believe VC is an optimal solution. But I believe it's "best" in the sense of that quote about capitalism -- "it's the worst system, except for everything else we've tried."

Incidentally, and I mean this quite seriously... If you think that one or more classes of founders are genuinely underrated, you *could* found a tech incubator to test the thesis. YC required some notoriety to found, but not a scary-huge amount of money. The odds are good that you could put something similar together for female founders if you wanted to.

(Not, like, "anybody could do that." But you, specifically, probably could.)
Camille, Destroyer of Worldsskamille on January 1st, 2014 09:47 pm (UTC)
I may someday be able to do that, but no, I cannot found an incubator for startups right now.

I really don't know what to tell you. I think you just vastly, vastly underestimate the impact of implicit sexism and racism. Do you really honestly think that the reason that so vanishingly few companies that are funded have minority founders is because those folks just don't exist at all? Or that they are all worse than the people that are getting funded? I work for a pair of successful female entrepreneurs, and while they won't name names, when they were raising funding, they had the sort of sexist incidents that would actually make your hair curl. And I can tell you having met many male founders, these ladies pretty much kick their asses up and down the street. There is some extreme overt sexism in VC. I actually don't think PG is an extreme overt sexist, but I think he has a ton of biases that he chooses not to address. I just recently heard a young woman who pitched to YC talk about how she was told "not to interrupt" when answering questions, when she observed several men do the same thing without comment.

VCs act like sheep. They all jumped on the bandwagon of incubators after the successes of YC. You are really giving them way too much credit, dude.
Camille, Destroyer of Worlds: pic#1093692skamille on January 1st, 2014 06:47 pm (UTC)
I think the real thing is that you either believe you have to be an AWESOME developer to create a good startup (in which case we should just be favoring older people), or you need to be qualified enough but also mostly about churning out tons of code (in which case discriminating against women is kind of absurd because most 13-to-17-year-olds are not actually doing the kind of practice that can't be caught up to in a year or so of disciplined study). But the implication is certainly not that a woman with fewer years of experience is just as good as a man with more years.

Personally, I think that most kids that start companies in their early 20s fail, or possibly hit on a good idea, and are then within 5 years essentially pushed out by the grownups that are needed to run the company. If you're PG, it doesn't matter if the kids get pushed out because you make your money, but the kids probably are cheaper and easier to manipulate. I don't really know where you get that PG is funding older entrepreneurs, the average age for his investments is 26. http://techcrunch.com/2010/07/30/ron-conway-paul-graham/
Noahangelbob on January 1st, 2014 07:15 pm (UTC)
Hm. I'd thought he was funding some 40+ entrepreneurs based on something he wrote, but I'm not finding the numbers to back that up. It's entirely possible he no longer does.

Edited at 2014-01-01 07:15 pm (UTC)