So, how about stepping back from the issue of race because that makes people stupid: if Silicon Valley (or a large chunk of the institutions which make it up) has a persistent bias against X, regardless of whether that is invidious or just a result of suboptimal processes, and that bias is unconnected to merit, then you should be able to profitably exploit it.
Silicon Valley has many biases, of varying levels of connection to actual merit. One is that it is a very chummy place: who you know is, historically, the first filter applied to you (there is an entire culture around intros) and some folks consider it to be the best available method of approximating merit. This suggests the existence of a strategy which repeatably beats the tar out of the Valley: invest in people before they are known by the powers-that-be, introduce them to the powers-that-be, make out like a bandit. You will probably end up rich enough to take up amusing hobbies like running message boards in your spare time.
It is entirely possible that Silicon Valley has a blind spot with respect to X, where X is a entrepreneurial demographic or a market or a business model or a geographic location or a personality type or a whatever. If X is unconnected to merit, then that suggests the existence of a strategy which will predictably produce crushing. I might even go so far as to say that absence of crushing is fairly persuasive circumstantial evidence about either the existence of the blind spot or about the connection to merit.
I think the article was a solidly not-stupid approach to the issue of race; I would like to see more discussion like it.
I don't think capitalism can solve all ills; even if there is an opportunity for, say, seed-stage investment for startups run by tech "minorities", there is a whole litany of necessities and problems which give it long odds of being successful. You have to find individuals who are both interested in the problem, and have the resources required to address it; they have to have the right recipe of skills; they, themselves, have to know certain people (that chumminess again); and, moreover, they have to be willing to throw significant capital at addressing a problem which might not exist in 5 years.
Or, to put it another way: I think there can be common agreement at this point that YC is not funding and accepting as many startups as they would like to, which implies that there is a strong market demand for another YC which could be as successful, or nearly as successful, as YC itself.
There are a lot of investment groups which have been, ahem, inspired by the YC model. I'm totally agnostic on what degree of success any particular one or the field in general will have, but in 2022, we'll have trivially checkable empirical evidence on whether there needed to be more seed stage funding than YC was doing for itself.
long odds of being successful
Do you realize that, if you believe the odds are long against this strategy succeeding, you have to be taking the position that this isn't a serious problem in the status quo? And vice versa? I mean, you're not going to give long odds against gravity or penicillin winning, right? They're going to ROFLstomp the alternatives. If diversity is not ROFLstomping the alternatives, then we did not have a diversity problem.
I realize that that is the conclusion if you believe that it is inevitable that a free market will capitalize on every opportunity.
I do not believe that this is so. I mentioned a few of the things which might prolong or even prevent such a thing from being successful even if there is sufficient opportunity for it.
I might be willing to change my position if there were a much greater number of people with the resources necessary to launch YC-like companies.
So I think we're coming at this in two different ways. It is possible that we're just philosophically irreconcilable on that.
I am skeptical, but willing to be convinced by evidence, that access to capital contributes more to success of software companies than talent. If, hypothetically, companies fail because of lacking access to capital but do not fail because of lacking access to talent, and the market selected for people who were fundable instead of people who were talented, I would not come to the conclusion that the free market has failed. I would come to the conclusion that the free market had allocated resources correctly to capital and away from talent (in the dystopian world where talent, apparently, doesn't really matter).
If you're betting that access to capital and social networks and execution risk trump diversity in investment outcomes, you're betting that access to capital and social networks and execution risk matter more than diversity.
I am skeptical, but willing to be convinced by evidence, that access to capital contributes more to success of software companies than talent.
One reason that the contrary belief is so persistent is that 'talent' is defined after the fact as that which was successful. If someone is talented but unsuccessful, you'll never hear of them. It's the dog that didn't bark. I believe that capital is a much better predictor of success than talent, as it is a much simpler, mathematically proven explanation for financial success and doesn't rely on a highly subjective judgment. I'm open to evidence to the contrary, but so far no one has produced any.
I think we are, too. Still, you're forcing me to think more carefully about my position on this, which I appreciate. I had to go outside and rake the leaves and think about what you said before replying.
So, I think you carried my argument a little bit further than I would have. For example, I don't think that access to capital contributes more to success than talent, nor do I think that talent doesn't really matter.
It's a bit more nuanced than that.
I understand that you're more attracted to data than supposition; unfortunately, I'm neither a researcher of sociology, nor prepared to locate data on "unsuccessful but extremely talented people" (which is where this is going) -- largely because I'm gobsmacked by the very thought of trying to dig up such data, especially historically.
The only thing that I have to fall back on is a particular series of reasoning:
- We assume that there are varying degrees of talent, i.e., "extremely talented" and, let's say, "normally" (or "moderately") talented.
- We assume that YC does not restrict itself only to teams of the most extremely talented. I think I can at least support this with pg's own past statements; he's far more likely to cite Sam Altman as an example of extreme talent than any other given YC graduate. I doubt pg would want to say that anyone in YC wasn't extremely talented, especially publicly, nor am I going to put those words in his mouth. However, it's clear from his own statements that YC has funded a range of differing talents, with Sam Altman on one end and many other people at another end.
- We assume that some of these YC graduates have been successful by nearly any measure. If these successful individuals are not Sam Altman, then we have an example of people who have been successful despite having less talent than someone else. (Again, I am very very carefully here not implying that they are not talented, or even that they are not extremely talented. This is not in any way to be taken as a criticism of anyone involved with YC in any way.)
- Given those, we have now shown a non-1:1 correspondence between talent and success.
- So: talent alone is not a predictor of success.
My position was not that talent doesn't matter, nor that access to capital matters more. My position was that access to capital matters. You can have people with identical levels of talent, and some will be successful, and some will not, based upon the resources that are available to them and any number of other semi-random influences.
I will readily agree that a person with an extreme amount of talent has better odds of succeeding than someone with less talent. They are more likely to possess the skills necessary to overcome a greater number of more serious challenges. But, those individuals are terribly rare in society; we would be trying to draw conclusions about sociology from a 1% of a 1% of a 1% of a 1%.
When discussing the greater general population, I think that environmental factors -- like access to funding and resources -- can greatly affect someone's success in life.
> If you're betting that access to capital and social networks and execution risk trump diversity in investment outcomes, you're betting that access to capital and social networks and execution risk matter more than diversity.
I think that the original article covered this:
"So, again, racial or gender diversity is not an end in itself. But we have to ask ourselves: if teams are consistently being put together with homogeneous demographics, what are the odds that they also will contain a diversity of perspectives? Shouldn’t we be worried that the same selection process that produces homogenous results in one area might be accidentally doing the same in the area that we care about (but that is harder to measure)?"
i.e., this is not about betting so much on a winning strategy, as wondering if perhaps we're collectively missing out on some opportunities.
Yeah, I scanned some of this conversation solely because I caught that it was an extremely civilized debate (rather than an emotionally heated argument). So amen to that. One of HN's finer moments.
Do you realize that, if you believe the odds are long against this strategy succeeding, you have to be taking the position that this isn't a serious problem in the status quo? And vice versa? I mean, you're not going to give long odds against gravity or penicillin winning, right? They're going to ROFLstomp the alternatives. If diversity is not ROFLstomping the alternatives, then we did not have a diversity problem.
Unfortunately, that is the crux of the problem. In the short-term diversity isn't really that important to the "majority". And diversity isn't going to ROFLstomp the alternatives. It isn't like there are a group of Blacks that are all Mark Zuckerbergs and Steve Jobs. White males are long shots for the most part as are Blacks.
The reason why diversity matters is more for society in the long-run. You want everyone to maximize their potential. And lastly, I think most people want others to have a fair shot at their full potential. People like to at least think they are fair.
So the diversity problem isn't about short-term ROI, but rather long-term maximization of human potential. Of course this is almost the total opposite of the argument proposed in the article.
Silicon Valley has many biases, of varying levels of connection to actual merit. One is that it is a very chummy place: who you know is, historically, the first filter applied to you (there is an entire culture around intros) and some folks consider it to be the best available method of approximating merit. This suggests the existence of a strategy which repeatably beats the tar out of the Valley: invest in people before they are known by the powers-that-be, introduce them to the powers-that-be, make out like a bandit. You will probably end up rich enough to take up amusing hobbies like running message boards in your spare time.
It is entirely possible that Silicon Valley has a blind spot with respect to X, where X is a entrepreneurial demographic or a market or a business model or a geographic location or a personality type or a whatever. If X is unconnected to merit, then that suggests the existence of a strategy which will predictably produce crushing. I might even go so far as to say that absence of crushing is fairly persuasive circumstantial evidence about either the existence of the blind spot or about the connection to merit.