5 ms·
I remember reading about a study which showed that, at least for certain companies, judging from first-decade performance, hiring based on interviews was less i
by andreyf 17y ago
I remember reading about a study which showed that, at least for certain companies, judging from first-decade performance, hiring based on interviews was less intelligent than hiring based purely on GPA [1]. The reasoning was that judgment during interviews is suceptible to various human biases to which GPA's are not.
At least part of PG's reasoning seems susceptible to similar biases: for example, whether or not he recognizes a username might have a lot to do with how memorable the username is. I wonder if sorting users by a subset of their upvotes - say, those received from news.YC alumni - would make more sense in "highlighting" intelligent applicants.
1. Would appreciate if anyone who recognizes this links it for me.
- pg 17y agoWe have lots of biases, but I don't think we're much influenced by how memorable usernames are. It's the comments that make a username memorable, not the name.
- andreyf 17y agoI don't think we're much influenced by how memorable usernames are Well, by definition, biases aren't intuitively obvious. But I'd love to hear explicit introspection about this in the essay: what are your biases, and how do you eliminate their affecting your decisions?
- pg 17y agoThat's a question we think about a lot, because bias basically equals things we judge wrong. I like energetic founders, for example, but one wouldn't call that a "bias," because energetic founders are good. So what you're really asking is what makes us pick wrong, which is something we're constantly thinking about. When I pick wrong, it's often because the founders seem really nice. That's probably my biggest weakness. The other YC partners have it too. It's much harder to get funded by YC if you're an asshole. But who knows, maybe this is actually an advantage. It means the atmos within YC is really good. Everyone tends to trust one another. So maybe it's a win overall, even if it decreases returns in individual cases. It's a complicated problem.
- grinich 17y agoIt's a complicated problem. Understated.
- messel 17y agoThe more eyes you have on it, the more varied the views, the weaker the bias becomes but also the more watered down the perspective. You have a solid "review board" size at 4-5. You guys & gal will minimize cognitive biases. I'm only commenting because I spent a few hours recently reading and writing about the topic, it's all new to me though (blog links are in HN).