A failure mode of these types of systems is that the people announcing events apply tags that are as general as possible to appeal to as wide variety of people as possible. So if you say you’re interested in, say, Jazz gigs, you might find notifications for things only tangentially related to Jazz. For instance, Blues or even Rock’n’Roll or whatever. Then it’s back to the old information overload.
A way to combat this might be to have an inverse rating system. So the more specific people’s tags are, and the fewer they are, the higher up they appear in your feed. This would incentivize people coming up with very specific tags that have a very high chance of being related to people’s interests.
This would incentivize people coming up with very specific tags that have a very high chance of being related to people’s interests.
It would also incentivize them to post the same stuff repeatedly, each time with a different tag. And use different descriptions, if you give lower rating to identical messages. And use different user accounts to post each message, if you give lower rating to multiple posts by the same user.
The duplication could be stopped if we had some kind of “attention economy”, where each person would receive e.g. one attention token every day, and they can use it with any message they think deserves attention. It would be possible to split the token for multiple messages (which most people will not see), collect multiple tokens for one extremely important message, and even adding your own tokens to support someone else’s message. The idea is that attention is a scarce resource, so it should be bought using a scarce money.
But even this system would probably fail, because… well, in real life the real money is scarce, advertising costs money, and yet we don’t exactly feel like advertising is the best source of information we would like to have. So even if we could introduce a world-wide attention economy that most people would use, there would gradually appear markets for buying your attention tokens for real money, and the real messages supported by a few real people could not compete with advertising supported by thousands of sold attention tokens.
However, these attention tokens in theory don’t have to be anonymous. They could be tracked to their original owner, even if the owner sells them. You could create your own filter and say that you only accept attention tokens by these specific people. (Presumably ones that don’t sell them.)
And now the last step is to somehow combine the scarcity of the attention tokens with the specificity of the tags given to a message. (Including the complicated details like: The same message can be supported by multiple people, but each of those people has a different opinion about the proper tags.)
I don’t see why that’s a problem; duplicate events could be detected by checking event locations and times. Events that seem to be occuring very near each other and at similar times could be ‘flagged’ for further investigation by moderators. The community could also help filter out other forms of abuse.
The first advice would be to make a web startup using some attention economy algorithms. However, it has the same problem as most of my ideas of startups: I have no idea how to make money out of it.
A failure mode of these types of systems is that the people announcing events apply tags that are as general as possible to appeal to as wide variety of people as possible. So if you say you’re interested in, say, Jazz gigs, you might find notifications for things only tangentially related to Jazz. For instance, Blues or even Rock’n’Roll or whatever. Then it’s back to the old information overload.
A way to combat this might be to have an inverse rating system. So the more specific people’s tags are, and the fewer they are, the higher up they appear in your feed. This would incentivize people coming up with very specific tags that have a very high chance of being related to people’s interests.
It would also incentivize them to post the same stuff repeatedly, each time with a different tag. And use different descriptions, if you give lower rating to identical messages. And use different user accounts to post each message, if you give lower rating to multiple posts by the same user.
The duplication could be stopped if we had some kind of “attention economy”, where each person would receive e.g. one attention token every day, and they can use it with any message they think deserves attention. It would be possible to split the token for multiple messages (which most people will not see), collect multiple tokens for one extremely important message, and even adding your own tokens to support someone else’s message. The idea is that attention is a scarce resource, so it should be bought using a scarce money.
But even this system would probably fail, because… well, in real life the real money is scarce, advertising costs money, and yet we don’t exactly feel like advertising is the best source of information we would like to have. So even if we could introduce a world-wide attention economy that most people would use, there would gradually appear markets for buying your attention tokens for real money, and the real messages supported by a few real people could not compete with advertising supported by thousands of sold attention tokens.
However, these attention tokens in theory don’t have to be anonymous. They could be tracked to their original owner, even if the owner sells them. You could create your own filter and say that you only accept attention tokens by these specific people. (Presumably ones that don’t sell them.)
And now the last step is to somehow combine the scarcity of the attention tokens with the specificity of the tags given to a message. (Including the complicated details like: The same message can be supported by multiple people, but each of those people has a different opinion about the proper tags.)
I don’t see why that’s a problem; duplicate events could be detected by checking event locations and times. Events that seem to be occuring very near each other and at similar times could be ‘flagged’ for further investigation by moderators. The community could also help filter out other forms of abuse.
Viliam, you’ve been rocking on my posts lately, can I hire you for personal counselor every now and then?
Sure. I’ll see what I can do for $200. :D
The first advice would be to make a web startup using some attention economy algorithms. However, it has the same problem as most of my ideas of startups: I have no idea how to make money out of it.