6 ms·

I definitely agree with this point. Discovery in the long-tail of content (without resorting to sort by new which is a nightmare) is a must solve problem in a i

I definitely agree with this point. Discovery in the long-tail of content (without resorting to sort by new which is a nightmare) is a must solve problem in a independent/decentralized web. I have a feeling the "answer" is "all of the above" options; the ability to choose what discovery/recommendation mechanisms to use. Being intentional about which lens your viewing your recommendations through. Being able to easily swap between lens. Allowing for corporation owned lens and community owned lens. Allowing yourself to look at an issue through the lens of an individual or institution that you trust. > people are really bad at linking to each other I wonder if the underlying point here is that the initial set of recommendations comes from small "communities". The readers of a certain blog suggest similar content to others in a non-spammy way. An independent music label links to artists that they enjoy listening to. I think there's missing infrastructure that makes it hard to form these communities though; any barrier to entry makes it less useful. I could see the "web3 ownership economy" thing helping here if it ever becomes a thing; creating this infrastructure as a side effect of the other stuff they want to accomplish. It'd be like if a subreddit was automatically created for every content creator, which could then be used to find recommendations of similar content.

jkhdigital 5y ago
I was reading a security paper today that made a killer observation about trust infrastructure: public trust systems will always succumb to problems of informational asymmetry (moral hazard, adverse selection) so responsible users must always rely on private trust assessments first. Same problem with information discovery—trust, reputation, it’s all the same social mechanism. Any public signaling will be degraded by information asymmetry. My point is that content discovery is a hard problem.
- zozbot234 5y ago
  > public trust systems will always succumb to problems of informational asymmetry (moral hazard, adverse selection) Ironically, this is something that the "smart contracts" powering some future development of Web 3.0 might eventually solve. By making it possible to verifiably put real-world money behind one's assertions and trust assessments, these problems are significantly reduced.
  - beej71 5y ago
    I think I'm misunderstanding. Wouldn't this just make us trust the people with the most money?
    - meheleventyone 5y ago
      Being charitable I think they mean some sort of betting site where people put money behind opinions and lose it if they are wrong or gain it if they are correct. Giving a sort of wisdom of the crowds. This of course is still subject to some authority who can say what is right or wrong on any particular issue so doesn’t escape the trust problem. Also the credulousness of the current crypto users makes me question how valuable any signal from it would actually be. The idea of merely trusting people because they’ve spent money is hilarious in the face of propaganda and advertising.
      - naasking 5y ago
        > Being charitable I think they mean some sort of betting site where people put money behind opinions and lose it if they are wrong or gain it if they are correct. Giving a sort of wisdom of the crowds. They're called prediction markets, and they already exist! * https://www.predictit.org https://www.predictit.org * https://www.metaculus.com/questions/ https://www.metaculus.com/questions/
        meheleventyone 5y ago
        Ahh, creeping dystopia!
        naasking 5y ago
        I dunno, I don't think it's necessarily a bad idea. Exploiting the wisdom of the crowds to make predictions could be a very good thing in fact. Unfortunately they fall afoul of gambling laws in some states/countries, so they're pretty limited right now. Scott does a decent overview here: https://astralcodexten.substack.com/p/metaculus-monday https://astralcodexten.substack.com/p/metaculus-monday

jkhdigital 5y ago

I was reading a security paper today that made a killer observation about trust infrastructure: public trust systems will always succumb to problems of informational asymmetry (moral hazard, adverse selection) so responsible users must always rely on private trust assessments first. Same problem with information discovery—trust, reputation, it’s all the same social mechanism. Any public signaling will be degraded by information asymmetry. My point is that content discovery is a hard problem.

zozbot234 5y ago
> public trust systems will always succumb to problems of informational asymmetry (moral hazard, adverse selection) Ironically, this is something that the "smart contracts" powering some future development of Web 3.0 might eventually solve. By making it possible to verifiably put real-world money behind one's assertions and trust assessments, these problems are significantly reduced.
- beej71 5y ago
  I think I'm misunderstanding. Wouldn't this just make us trust the people with the most money?
  - meheleventyone 5y ago
    Being charitable I think they mean some sort of betting site where people put money behind opinions and lose it if they are wrong or gain it if they are correct. Giving a sort of wisdom of the crowds. This of course is still subject to some authority who can say what is right or wrong on any particular issue so doesn’t escape the trust problem. Also the credulousness of the current crypto users makes me question how valuable any signal from it would actually be. The idea of merely trusting people because they’ve spent money is hilarious in the face of propaganda and advertising.
    - naasking 5y ago
      > Being charitable I think they mean some sort of betting site where people put money behind opinions and lose it if they are wrong or gain it if they are correct. Giving a sort of wisdom of the crowds. They're called prediction markets, and they already exist! * https://www.predictit.org https://www.predictit.org * https://www.metaculus.com/questions/ https://www.metaculus.com/questions/
      - meheleventyone 5y ago
        Ahh, creeping dystopia!
        naasking 5y ago
        I dunno, I don't think it's necessarily a bad idea. Exploiting the wisdom of the crowds to make predictions could be a very good thing in fact. Unfortunately they fall afoul of gambling laws in some states/countries, so they're pretty limited right now. Scott does a decent overview here: https://astralcodexten.substack.com/p/metaculus-monday https://astralcodexten.substack.com/p/metaculus-monday