4 ms·

The PeopleRank (or Whuffie or Karma or Facebook Points or eBay rating or whatever) is coming, no doubt. This is one of those major shifts to society that one d

The PeopleRank (or Whuffie or Karma or Facebook Points or eBay rating or whatever) is coming, no doubt. This is one of those major shifts to society that one day will just be there and change everything. Are you ready for it? Are we?

patio11 17y ago
I don't even think it is that huge a shift from what we already accept, although it will be a quantum leap in effectiveness. I think a small team of folks with CS backgrounds plus an insurance company's claim data (or a bank's loss data) for a year plus a Hadoop cluster would make FICO obsolete, at least until they were regulated out of existence for being too darned efficient at what they do. (The basic theory is "Birds of a feather flock together": if most of your friends are losers, it is probably net-beneficial to the business to assume you are a loser, if most of our friends are solid debt-paying responsible citizens, it is probably net-beneficial to the business to assume you are a solid debt-paying responsible citizen.) I think this would be so effective (prices risk appropriately, far cheaper to maintain than the current reporting infrastructure since folks helpfully update the social graph for you) you'd almost have to illegalize it. (Let me hum a few bars: "disparate impact".)
- barrkel 17y ago
  Well, what you describe is all but the very definition of prejudice. It would discriminate directly against whole social classes of people irrespective of their individual behaviour, on the basis of group behaviour. From what you describe, in a private education system built around loans it would have kicked me out at 15 or so. On the other hand, such an inefficient classification mechanism ought to create sufficient surplus for profit for those specialized enough to look for it. But there are other systemic inefficiencies that might prevent those niches from being profitable enough.
  - Confusion 17y ago
    Someone that has relatively many friends that play football is probably someone that likes, and plays, football. Categorizing him as such is neither prejudiced nor discriminatory: it's a reasonable extrapolation. The same holds for more sensitive facts about someone: someone with many gay friends is more likely than average to be gay. There already are insurance companies that only insure people with a university degree. Their rates are lower than those of other companies. They aren't prejudiced and they do not discriminate: they estimate risk based on solid information about a person and the mere fact that they can stay in business suggests there is indeed a good correlation between their estimates risk and the actual risk. If they could determine who has many university educated friends, then they would perhaps want to extend their service to those people, because they have determined that those people share the characteristics of university educated people that cause the insured risks to be smaller. (Assuming these reasons that are independent of the reasons for not doing/finishing a university education, etc.)
    - barrkel 17y ago
      Insurance is explicitly about pooling risk, and treating individuals along group lines. If the insurance companies could isolate, to the individual level, what the exact risk is for that person, it wouldn't be insurance any more.
      - gizmo 17y ago
        Nope, it doesn't work like that. Insurance is about reducing variance. If I know the chance of crashing my car and losing a leg is exactly 1 / 50000 (per year), and if I can insure against that at roughly that rate, I would still get insurance, because $20 bucks a year doesn't affect my life, but a million will help a lot when I do lose that leg. In this case risk is still pooled and the insurance company can still function because the misfortune of the individual is paid for by the more fortunate. There is still a need for insurance companies exist in a world with perfect information. Of course the insurance rates would shift. For example, when in a large group of people a few have a high chance of an expensive disease the insurance company wouldn't give the same rate to everybody. You'd still want to get insurance, though.
    - 17y ago
  - anamax 17y ago
    > Well, what you describe is all but the very definition of prejudice. Actually, it's not. When insurance is correctly priced, the payout for a group is close to its premiums, no matter how the group is defined. (Yes, including factors that weren't considered by the insurance company.) If you don't price that way, some group is getting a subsidy and another group is paying one. An insurance company that figures out how to identify the latter can offer its members better rates than a company that doesn't. Or, if the smarter company can figure out how to avoid the former, it can have higher profits (which attracts competitors). Insurance companies like to make money. People like to pay the minimum for their insurance. These two facts have economic consequences.
  - byrneseyeview 17y ago
    That's the point. If the prediction is accurate, that's net beneficial: irresponsible people will get fewer loans, and responsible people will get more loans. It sounds like you haven't quite decided whether this would be less accurate than FICO (maybe! But there's no reason to be outraged that some random person's prejudice is costing them money) or more accurate than FICO (in which case your argument seems to be "Yes, this kind of prejudice is accurate--or, at least, more accurate than the alternative. But that makes me feel bad, so misallocate capital and make more bad loans.") The world economy just went through a couple very bad years caused by making wishful-thinking loans instead of unpleasant-truth loans. I hope that doesn't have to happen again.
cgs 17y ago
Reminds me of the "darknet" in Daniel Suarez's Daemon, a sort of augmented reality. People could be virtually downvoted/upvoted and carried their reputation scores around floating above their heads.

patio11 17y ago

I don't even think it is that huge a shift from what we already accept, although it will be a quantum leap in effectiveness. I think a small team of folks with CS backgrounds plus an insurance company's claim data (or a bank's loss data) for a year plus a Hadoop cluster would make FICO obsolete, at least until they were regulated out of existence for being too darned efficient at what they do. (The basic theory is "Birds of a feather flock together": if most of your friends are losers, it is probably net-beneficial to the business to assume you are a loser, if most of our friends are solid debt-paying responsible citizens, it is probably net-beneficial to the business to assume you are a solid debt-paying responsible citizen.) I think this would be so effective (prices risk appropriately, far cheaper to maintain than the current reporting infrastructure since folks helpfully update the social graph for you) you'd almost have to illegalize it. (Let me hum a few bars: "disparate impact".)

barrkel 17y ago
Well, what you describe is all but the very definition of prejudice. It would discriminate directly against whole social classes of people irrespective of their individual behaviour, on the basis of group behaviour. From what you describe, in a private education system built around loans it would have kicked me out at 15 or so. On the other hand, such an inefficient classification mechanism ought to create sufficient surplus for profit for those specialized enough to look for it. But there are other systemic inefficiencies that might prevent those niches from being profitable enough.
- Confusion 17y ago
  Someone that has relatively many friends that play football is probably someone that likes, and plays, football. Categorizing him as such is neither prejudiced nor discriminatory: it's a reasonable extrapolation. The same holds for more sensitive facts about someone: someone with many gay friends is more likely than average to be gay. There already are insurance companies that only insure people with a university degree. Their rates are lower than those of other companies. They aren't prejudiced and they do not discriminate: they estimate risk based on solid information about a person and the mere fact that they can stay in business suggests there is indeed a good correlation between their estimates risk and the actual risk. If they could determine who has many university educated friends, then they would perhaps want to extend their service to those people, because they have determined that those people share the characteristics of university educated people that cause the insured risks to be smaller. (Assuming these reasons that are independent of the reasons for not doing/finishing a university education, etc.)
  - barrkel 17y ago
    Insurance is explicitly about pooling risk, and treating individuals along group lines. If the insurance companies could isolate, to the individual level, what the exact risk is for that person, it wouldn't be insurance any more.
    - gizmo 17y ago
      Nope, it doesn't work like that. Insurance is about reducing variance. If I know the chance of crashing my car and losing a leg is exactly 1 / 50000 (per year), and if I can insure against that at roughly that rate, I would still get insurance, because $20 bucks a year doesn't affect my life, but a million will help a lot when I do lose that leg. In this case risk is still pooled and the insurance company can still function because the misfortune of the individual is paid for by the more fortunate. There is still a need for insurance companies exist in a world with perfect information. Of course the insurance rates would shift. For example, when in a large group of people a few have a high chance of an expensive disease the insurance company wouldn't give the same rate to everybody. You'd still want to get insurance, though.
  - 17y ago
- anamax 17y ago
  > Well, what you describe is all but the very definition of prejudice. Actually, it's not. When insurance is correctly priced, the payout for a group is close to its premiums, no matter how the group is defined. (Yes, including factors that weren't considered by the insurance company.) If you don't price that way, some group is getting a subsidy and another group is paying one. An insurance company that figures out how to identify the latter can offer its members better rates than a company that doesn't. Or, if the smarter company can figure out how to avoid the former, it can have higher profits (which attracts competitors). Insurance companies like to make money. People like to pay the minimum for their insurance. These two facts have economic consequences.
- byrneseyeview 17y ago
  That's the point. If the prediction is accurate, that's net beneficial: irresponsible people will get fewer loans, and responsible people will get more loans. It sounds like you haven't quite decided whether this would be less accurate than FICO (maybe! But there's no reason to be outraged that some random person's prejudice is costing them money) or more accurate than FICO (in which case your argument seems to be "Yes, this kind of prejudice is accurate--or, at least, more accurate than the alternative. But that makes me feel bad, so misallocate capital and make more bad loans.") The world economy just went through a couple very bad years caused by making wishful-thinking loans instead of unpleasant-truth loans. I hope that doesn't have to happen again.

cgs 17y ago

Reminds me of the "darknet" in Daniel Suarez's Daemon, a sort of augmented reality. People could be virtually downvoted/upvoted and carried their reputation scores around floating above their heads.