In my last post I suggested that we prefer institutions of this form: Masses recognize elites, who oversee experts, who pick details. However, our existing methods for doing that first step, masses recognizing elites
But the app could give elites much more say over who is elite, and satisfy that elite desire to say in control of who is elite. And elites might see the app as doing this all more accurately, with less noise than in their usual informal processes.
My presumption is that this is self-defeating - in some senses masses choose elites but part of being elite is telling the masses who is elite, who they should say is elite, who they have said is elite, and so on, and for judgments to be made in elite fashion via opaque elite systems. So like all other similar systems, expect elites to reject the system, and for masses to therefore reject it as well.
In a world of tribes, I expect most respondents to say, "My tribe member is better." So the status poll mostly just tells me who belongs to the tribe of folks replying to the poll. What am I missing?
Yes, if you just do it as 'A yay or boo B', you will degenerate into popularity and discover that Kim Kardashian is the 'highest status person in the world' or something. You would need to make it contextual and start building up graphs of status: think more sophisticated proposals for ranking like PageRank or Scott Aaronson's eigen democracy or Avogadro, which try to propagate metrics. To take your Tor example, I was surprised to learn recently (not following literature matters much these days) that Jenna Bush has become a huge success in the book club world (https://archive.vn/DIPwd); if you asked a random person about "who is a high-status book editor?", they will have no idea about most of the suggestions, but they will say they trust Jenna Bush on 'books' (and not, say, rocket surgery), and Jenna Bush surely has many and well-informed opinions on the closely connected nodes to 'books' like the status ranking of 'book editors'.
One question is whether users would actually enjoy spending large amounts of time deciding between A and B. I could see a lot of people finding it off-putting and not wanting to do it. And then it leads to an issue where the rankings are coming from a biased sample of the population, ie the people who are actually motivated to spend hours deciding between A and B.
A second question is how you would find the right A and B for X to vote on. This was mentioned in the post, but I want to call it out. As a user if I kept getting A's and B's who I didn't know well enough to feel qualified to judge, I'd hit the "skip" button (assuming there is one). But after hitting "skip" over and over, I'd eventually get bored and go somewhere else.
A third question is how to bootstrap this. In the early days when there isn't a lot of usage, why would people be motivated to use it?
A fourth question is whether people would actually want to spend time looking through status estimates. Thinking about myself, if this was popular I could see myself screwing around a little bit looking at the estimates for a handful of people I'm curious about but not the sort of thing I'd spend hours and hours on each week, let alone day. Relatedly, there's the question of how this would be monetized. If this was free and monetized via ads, it'd have to be super popular, but I question whether it's the sort of thing people would spend hours and hours of time on each week. If it was paid, who would care enough to pay? I'm skeptical that people would pay >$100 to view different statuses.
With your seventh issue - you can always reduce X, Y, ... ranking A, B, .. to triples as long as there is at least one X, and two A, Bs, but, it'd likely be possible to get more info from not treating it as separate pairs.
You might want to consider using a difference score rather than just a ranking for each pair to get more info out of each 'vote' or whatever it is called. You wouldn't necessarily treat the score linearly.
I think you've highlighted a lot of fine tuning issues there and maybe missed some basic things like people who aren't real getting in the system, people getting in the system multiple times, someone starting right and left wing versions of the app giving opposite rankings to people.
Trouble is . . . status isn't a single axis. One of the features of modern society is precisely the contingent and situational nature of status.
Example: the publisher at Tor Books is one of the most influential and high-status people in science fiction. But outside that community, he's just another mope with a desk job. How can you meaningfully measure whether he's higher status than a first-rate college football player? If you poll the entire population, he'll probably wind up well below.
This is probably why so many people comment on how "trashy" celebrities are nowadays -- in the absence of a Mrs. Astor's 400 to set the social status scale, we just go by how famous people are, and one can be famous for just about anything.
And in many areas there are opposed, highly bimodal status pyramids. How high is Hilary Clinton among Trump voters? Or vice versa? How do you decide which one "deserves" higher status?
If rankings are visible to each ranked person (even if otherwise secret), and we rank everybody, the rankings might incent better public behavior (to get the rankings up).
Maybe useful for incenting better behavior on, for example, social media. Which would be a huge win, independent of institutional design.
1 - We could keep the results secret for most people. Those who want their results public could allow that. Or maybe those who want to run for political office would have to.
2 - We could have multiple kinds of "respect" ratings. I really respect Albert Einstein's abilities as a physicist, and William Shakespeare's abilities as a writer. I'm not sure I want either one making policy.
If each person has multiple respect ratings, we no longer have a strict ranking of social order.
IMO, light illegibility of status is very important because having a precise measure of your position in the social order is very harsh. Venkatesh Rao has a great discussion of this in his book on the Office: that a ranking is a catastrophic idea, and an illusion that everyone is special in some way is essential
"Defer to" is too strong a phrase. It suggests that if person A says Y, I can stop thinking entirely about Y and just accept what A said. There are some people who often think in these terms, but it's not often rational to just ignore all other sources of information besides what an expert said.
An important question regarding this business proposal is finding out what the "value proposition hypothesis is". Something like "people want to use a new platform where they can go and find more information about who is high-status". First recommended step in a startup is usually to test this hypothesis in the cheapest way possible, so the first prototype would be easy, simple-to-obtain information X (some information related to status) about some subset of people and then we can test whether "people want to go to scrappy website to get information X about those people in a more convenient way".
I think this concept needs an MVP and having a "general status tracker" may be too high of a bar for a first version.
A narrow version such as "Twitter Prestige" would probably be useful already, so why not start with that? If that booms then it can become more general later.
Another advantage of that as an MVP is that it leverages an existing platform for early growth. This is usually quite important for startups (e.g. Paypal leveraging Ebay) -- books such as Blitzcaling discuss advantages of that.
Also leveraging social networks means you can start with lots of public data (who follows who, public likes and comments) bypassing a cold start problem.
But the app could give elites much more say over who is elite, and satisfy that elite desire to say in control of who is elite. And elites might see the app as doing this all more accurately, with less noise than in their usual informal processes.
My presumption is that this is self-defeating - in some senses masses choose elites but part of being elite is telling the masses who is elite, who they should say is elite, who they have said is elite, and so on, and for judgments to be made in elite fashion via opaque elite systems. So like all other similar systems, expect elites to reject the system, and for masses to therefore reject it as well.
In a world of tribes, I expect most respondents to say, "My tribe member is better." So the status poll mostly just tells me who belongs to the tribe of folks replying to the poll. What am I missing?
Yes, if you just do it as 'A yay or boo B', you will degenerate into popularity and discover that Kim Kardashian is the 'highest status person in the world' or something. You would need to make it contextual and start building up graphs of status: think more sophisticated proposals for ranking like PageRank or Scott Aaronson's eigen democracy or Avogadro, which try to propagate metrics. To take your Tor example, I was surprised to learn recently (not following literature matters much these days) that Jenna Bush has become a huge success in the book club world (https://archive.vn/DIPwd); if you asked a random person about "who is a high-status book editor?", they will have no idea about most of the suggestions, but they will say they trust Jenna Bush on 'books' (and not, say, rocket surgery), and Jenna Bush surely has many and well-informed opinions on the closely connected nodes to 'books' like the status ranking of 'book editors'.
One question is whether users would actually enjoy spending large amounts of time deciding between A and B. I could see a lot of people finding it off-putting and not wanting to do it. And then it leads to an issue where the rankings are coming from a biased sample of the population, ie the people who are actually motivated to spend hours deciding between A and B.
A second question is how you would find the right A and B for X to vote on. This was mentioned in the post, but I want to call it out. As a user if I kept getting A's and B's who I didn't know well enough to feel qualified to judge, I'd hit the "skip" button (assuming there is one). But after hitting "skip" over and over, I'd eventually get bored and go somewhere else.
A third question is how to bootstrap this. In the early days when there isn't a lot of usage, why would people be motivated to use it?
A fourth question is whether people would actually want to spend time looking through status estimates. Thinking about myself, if this was popular I could see myself screwing around a little bit looking at the estimates for a handful of people I'm curious about but not the sort of thing I'd spend hours and hours on each week, let alone day. Relatedly, there's the question of how this would be monetized. If this was free and monetized via ads, it'd have to be super popular, but I question whether it's the sort of thing people would spend hours and hours of time on each week. If it was paid, who would care enough to pay? I'm skeptical that people would pay >$100 to view different statuses.
With your seventh issue - you can always reduce X, Y, ... ranking A, B, .. to triples as long as there is at least one X, and two A, Bs, but, it'd likely be possible to get more info from not treating it as separate pairs.
You might want to consider using a difference score rather than just a ranking for each pair to get more info out of each 'vote' or whatever it is called. You wouldn't necessarily treat the score linearly.
I think you've highlighted a lot of fine tuning issues there and maybe missed some basic things like people who aren't real getting in the system, people getting in the system multiple times, someone starting right and left wing versions of the app giving opposite rankings to people.
Especially if they aren't real people.
Trouble is . . . status isn't a single axis. One of the features of modern society is precisely the contingent and situational nature of status.
Example: the publisher at Tor Books is one of the most influential and high-status people in science fiction. But outside that community, he's just another mope with a desk job. How can you meaningfully measure whether he's higher status than a first-rate college football player? If you poll the entire population, he'll probably wind up well below.
This is probably why so many people comment on how "trashy" celebrities are nowadays -- in the absence of a Mrs. Astor's 400 to set the social status scale, we just go by how famous people are, and one can be famous for just about anything.
And in many areas there are opposed, highly bimodal status pyramids. How high is Hilary Clinton among Trump voters? Or vice versa? How do you decide which one "deserves" higher status?
If rankings are visible to each ranked person (even if otherwise secret), and we rank everybody, the rankings might incent better public behavior (to get the rankings up).
Maybe useful for incenting better behavior on, for example, social media. Which would be a huge win, independent of institutional design.
1 - We could keep the results secret for most people. Those who want their results public could allow that. Or maybe those who want to run for political office would have to.
2 - We could have multiple kinds of "respect" ratings. I really respect Albert Einstein's abilities as a physicist, and William Shakespeare's abilities as a writer. I'm not sure I want either one making policy.
If each person has multiple respect ratings, we no longer have a strict ranking of social order.
Thanks!
IMO, light illegibility of status is very important because having a precise measure of your position in the social order is very harsh. Venkatesh Rao has a great discussion of this in his book on the Office: that a ranking is a catastrophic idea, and an illusion that everyone is special in some way is essential
Sure, I have no problem with starting with a smaller project. But I still wanted to think about where we might want to go eventually.
"Defer to" is too strong a phrase. It suggests that if person A says Y, I can stop thinking entirely about Y and just accept what A said. There are some people who often think in these terms, but it's not often rational to just ignore all other sources of information besides what an expert said.
An important question regarding this business proposal is finding out what the "value proposition hypothesis is". Something like "people want to use a new platform where they can go and find more information about who is high-status". First recommended step in a startup is usually to test this hypothesis in the cheapest way possible, so the first prototype would be easy, simple-to-obtain information X (some information related to status) about some subset of people and then we can test whether "people want to go to scrappy website to get information X about those people in a more convenient way".
I think this concept needs an MVP and having a "general status tracker" may be too high of a bar for a first version.
A narrow version such as "Twitter Prestige" would probably be useful already, so why not start with that? If that booms then it can become more general later.
Another advantage of that as an MVP is that it leverages an existing platform for early growth. This is usually quite important for startups (e.g. Paypal leveraging Ebay) -- books such as Blitzcaling discuss advantages of that.
Also leveraging social networks means you can start with lots of public data (who follows who, public likes and comments) bypassing a cold start problem.