If you asked college admissions officers in the 1950's what kind of student would be most likely to succeed, they would probably describe someone who was white,

If you asked college admissions officers in the 1950's what kind of student would be most likely to succeed, they would probably describe someone who was white, male, and came from a middle-to-upper-class background. Not because those kinds of students are inherently superior, but only because the vast majority of college students at the time came from this very limited demographic. There's the same danger with VC. It's probably/hopefully an unconscious bias, but it's very tempting to want to invest in a "type" of founder that has succeeded often in the past - Young, white, male, Harvard/Stanford dropout, wears a hoodie, etc. Every VC wants to invest in the next Bill Gates or Mark Zuckerberg, but the more we fixate on past successful archetypes, the more likely we are to (unconsciously) dismiss other possibilities. It's my same gripe with the job interview process. Obviously every company wants to hire "the smartest, hardest-working people who are a good cultural fit". But how do you measure that in a way that's both accurate and efficient/scalable? You usually can't, which causes you to use certain heuristics which optimize for efficiency/scalability at the expense of considering everyone who might be a good candidate. For example, the heuristic of using internal referrals for hiring works, because if you're a company like Google where everyone is smart, then smart people are likely to know other smart people, and since you're referring friends, it's likely they will be a good cultural fit. But obviously, in the long run internal referrals will lead to a very homogeneous culture, which can be good (everyone gets along) and bad (lack of diversity).