4 ms·

I think the key to achieving this is building Web of Trust on top of the fact checker relationships and connect them to the users, so that you can see exactly h

I think the key to achieving this is building Web of Trust on top of the fact checker relationships and connect them to the users, so that you can see exactly how you are connected, and whether you can trust the individuals. Today, the sources of whatever information you read, even from the most credible sources, are obscured. Whether it is CNN or a peer-reviewed journal article, the most information you will get about the author is a name, a picture, and maybe a bit of their publicly available data. With Web of Trust, the level of trust you should assign to an author, and by extension the information, can be instantaneously computed by adding up your own trust references. If I assign my trust responsibly, and you assign your trust to me, then everyone I trust is by extension trustworthy, and the information metadata can reflect this. If you don't think I assign my trust responsibly, then you can remove me from your trust circle, and my assigned trust will become irrelevant. It's not a perfect system, it's vulnerable to attacks, it requires eternal vigilance, but it's the best we've got going for us today, IMO.

awinter-py 6y ago
have you used a web of trust system you liked, that works well w/ an existing social platform? I'd try one if you have a rec. fwiw I wrote one of these and mine sucked
- jancsika 6y ago
  I'm curious-- did you build it as a decentralized/federated/distributed piece of software?
  - awinter-py 6y ago
    nope -- it was a forum, the web-of-trust feature only operated over the users and content inside the forum
    - forgotmypw17 6y ago
      Could you please write a little more about your forum? I'm also working on a forum system that includes basic WoT features, and would be grateful for anything you share. (Prototype link is in my profile if you want to play with the current prototype.)
      - jancsika 6y ago
        I second that! I'm glad to hear OP at least attempted WoT inside a centralized forum. I think one of the problems with WoT is that many projects hard-couple it to a distributed approach to software that utilizes it. Since OP didn't take that approach I'd be curious to know what didn't work.
        awinter-py 6y ago
        it wasn't a design decision to link the WoT to a single forum -- I was just playing around and didn't want to go through the hoops of integrated with an external social network Mine is here -- https://github.com/abe-winter/trustgator https://github.com/abe-winter/trustgator
- forgotmypw17 6y ago
  It's still very much a pre-MVP prototype, but try the link in my profile.
Dopameaner 6y ago
This system reminds me of a sophisticated Facebook news feed. Though a major problem is the news sources that are biased at the source. We might need a transparency fact checker that compares between sources.
toomim 6y ago
I'm working on building this. Email me if interested! toomim@gmail.com
- forgotmypw17 6y ago
  I sent you an email, did you receive it?
zby 6y ago
Here is a design proposal for such a system with an interesting twist: https://www.lesswrong.com/posts/ZdtFBCtixqay5aoWF/design-thoughts-for-building-a-better-kind-of-social-space https://www.lesswrong.com/posts/ZdtFBCtixqay5aoWF/design-tho...
AnimalMuppet 6y ago
I'll go further. We need at least two webs of trust. Would you trust a Democrat to fact-check Trump's claims of election fraud? Would you trust a Republican? In fact, you should trust the Democrat if they say the claims are true, and the Republican if they say the claims are false. Ideally, we'd get one set of fact-checkers that both sides trust. In practice, that may not be possible. But at least, if we have separate fact-checkers for each side, if they agree, then we have a fairly good basis for suspecting that they're right.
- loopz 6y ago
  For most people outside US, like over 7 billion people, both of these would be invalid webs of trust.
  - forgotmypw17 6y ago
    I wouldn't trust anyone who uses partisan lines to assign trust. And, in a transparent WoT system, I would be able to see them doing this with an easy SQL query. I would then remove any #trust in this individual, and this particular WoT's trust level would be worthless (in my graph.)
  - redis_mlc 6y ago
    > For most people outside US, like over 7 billion people, both of these would be invalid webs of trust. Why and how does their opinion matter? It doesn't.
- forgotmypw17 6y ago
  That's a good point. I think more than one Web is a great idea and desirable. As far as your questions: I would trust some people who identify as Republican, but not others. I would trust some people who identify as Democrat, but not others. Out of the two sets, I would trust the ones whom I personally know, and not anyone else, no matter who they say they are. In the following scenario, I would trust both a Democrat and a Republican fact-checking Trump's claims of election fraud; [Me] #trust [My friend Alice, whom I trust to assign trust responsibly] #trust [Alices's friend Bob, whom Alice trusts] #trust [...] #trust [Democrat fact-checker] #approve fact-checking report. Depending on the Bacon number and how many other similar connections there are, I would trust them more or less. If there are no connections, then I wouldn't trust them beyond the distance I can propel their supposedly human body under my own power in Earth's gravity.
alexisread 6y ago
I'll drop https://spritelyproject.org/ https://spritelyproject.org/ here. Having said that, there is likely to be an issue of influencers ie. Those celebrities that you follow as they're interesting / exciting etc but may be completely uninformed on say green issues, which then have an undue influence on their followers. Worse, many of them are sponsored and corruptible that way - we're talking about boiling frogs both for the influencer and followers basically. The other issue is that a Web of trust has a tendency to echo chamber. I think the actual hardest thing to do is to cross-polinate networks in a voluntary way, that I don't know how to do yet.

awinter-py 6y ago

have you used a web of trust system you liked, that works well w/ an existing social platform? I'd try one if you have a rec. fwiw I wrote one of these and mine sucked

jancsika 6y ago
I'm curious-- did you build it as a decentralized/federated/distributed piece of software?
- awinter-py 6y ago
  nope -- it was a forum, the web-of-trust feature only operated over the users and content inside the forum
  - forgotmypw17 6y ago
    Could you please write a little more about your forum? I'm also working on a forum system that includes basic WoT features, and would be grateful for anything you share. (Prototype link is in my profile if you want to play with the current prototype.)
    - jancsika 6y ago
      I second that! I'm glad to hear OP at least attempted WoT inside a centralized forum. I think one of the problems with WoT is that many projects hard-couple it to a distributed approach to software that utilizes it. Since OP didn't take that approach I'd be curious to know what didn't work.
      - awinter-py 6y ago
        it wasn't a design decision to link the WoT to a single forum -- I was just playing around and didn't want to go through the hoops of integrated with an external social network Mine is here -- https://github.com/abe-winter/trustgator https://github.com/abe-winter/trustgator
forgotmypw17 6y ago
It's still very much a pre-MVP prototype, but try the link in my profile.

Dopameaner 6y ago

This system reminds me of a sophisticated Facebook news feed. Though a major problem is the news sources that are biased at the source. We might need a transparency fact checker that compares between sources.

toomim 6y ago

I'm working on building this. Email me if interested! toomim@gmail.com

forgotmypw17 6y ago
I sent you an email, did you receive it?

zby 6y ago

Here is a design proposal for such a system with an interesting twist: https://www.lesswrong.com/posts/ZdtFBCtixqay5aoWF/design-thoughts-for-building-a-better-kind-of-social-space https://www.lesswrong.com/posts/ZdtFBCtixqay5aoWF/design-tho...

AnimalMuppet 6y ago

I'll go further. We need at least two webs of trust. Would you trust a Democrat to fact-check Trump's claims of election fraud? Would you trust a Republican? In fact, you should trust the Democrat if they say the claims are true, and the Republican if they say the claims are false. Ideally, we'd get one set of fact-checkers that both sides trust. In practice, that may not be possible. But at least, if we have separate fact-checkers for each side, if they agree, then we have a fairly good basis for suspecting that they're right.

loopz 6y ago
For most people outside US, like over 7 billion people, both of these would be invalid webs of trust.
- forgotmypw17 6y ago
  I wouldn't trust anyone who uses partisan lines to assign trust. And, in a transparent WoT system, I would be able to see them doing this with an easy SQL query. I would then remove any #trust in this individual, and this particular WoT's trust level would be worthless (in my graph.)
- redis_mlc 6y ago
  > For most people outside US, like over 7 billion people, both of these would be invalid webs of trust. Why and how does their opinion matter? It doesn't.
forgotmypw17 6y ago
That's a good point. I think more than one Web is a great idea and desirable. As far as your questions: I would trust some people who identify as Republican, but not others. I would trust some people who identify as Democrat, but not others. Out of the two sets, I would trust the ones whom I personally know, and not anyone else, no matter who they say they are. In the following scenario, I would trust both a Democrat and a Republican fact-checking Trump's claims of election fraud; [Me] #trust [My friend Alice, whom I trust to assign trust responsibly] #trust [Alices's friend Bob, whom Alice trusts] #trust [...] #trust [Democrat fact-checker] #approve fact-checking report. Depending on the Bacon number and how many other similar connections there are, I would trust them more or less. If there are no connections, then I wouldn't trust them beyond the distance I can propel their supposedly human body under my own power in Earth's gravity.

alexisread 6y ago

I'll drop https://spritelyproject.org/ https://spritelyproject.org/ here. Having said that, there is likely to be an issue of influencers ie. Those celebrities that you follow as they're interesting / exciting etc but may be completely uninformed on say green issues, which then have an undue influence on their followers. Worse, many of them are sponsored and corruptible that way - we're talking about boiling frogs both for the influencer and followers basically. The other issue is that a Web of trust has a tendency to echo chamber. I think the actual hardest thing to do is to cross-polinate networks in a voluntary way, that I don't know how to do yet.