Agalmos -- an information measure for agalmic economics

"See! You get ahead by giving! Get with the program! Only the generous survive!

Charles Stross, in Accelerando

Motivation

  1. This is the post-scarcity economy. Right now. Except that brain cycles are still scarce
  2. Information abounds. Attention does not. And who do we pay attention to when everyone is giving information for free?
  3. We live in John Perry Barlow's independent cyberspace. We get ahead by giving. But how to sort out this chaos?
  4. Graph density measures like PageRank have been a popular and efficient method. People vote with their feet and information gets sorted out by votes.
  5. But are all votes the same? Is this just one more rant you find at reddit or digg or whatever, or is it something you pay attention to?
  6. Sorting out the chaos -- and hence giving rise to reputation markets where you can find out who matters and who doesn't in an attention-scarce economy -- is a matter of finding out how much attention you are receiving.
  7. One way to do this -- with all sorts of good mathematical properties I shall not prove -- is information entropy. How much do you deviate from the random noise of hordes of people absorbing large amounts of attention?
  8. We need to make this work -- otherwise, there will always be a scarce market of people providing "solutions" and sorting out the chaos into scarce non-chaos.

Mechanics

  1. First of all, defining a vote is a complicated business.
  2. Agalmos (please suggest a suitable plural form) was first conceived for the Twitter platform, where people can "follow" other people's updates, and the number of users one is following is assumed to be inversely proportional to the attention each feed is receiving. Then we use information entropy. Look up Wikipedia.

    Behold the Twitter/RSS/"subscribing model" formula to determine agalmos:


    (Worry not, the image is stolen from the blog I used to run.)
    where i=1,2,.. indexes the number of subscribers one has and mi is the number of feeds that person subscribes to.
  3. Is this adaptable? mi can be the number of clicks or upvotes to a given link. mi can be infinitely handed out whuffie, and the formula will sort it out.

How to do it

  1. This could be done on an entirely new platform. It could be sweet, a true reputation market.
  2. It probably will work for the best if we pick up a database from an existing service with an API. Twitter won't do because the API won't open up how many people are followed by a given follower. Reddit wouldn't work because there isn't a register of upvotes.
  3. In Reddit, Slashdot, etc. there's a database of each users' comments and the karma it receives. But you still can't figure out who's giving out the karma to see how much the karma's worth.
  4. How this will work is an open question. I'm throwing the seeds. Please discuss it everywhere you can.