4 ms·

I honestly hope for a framework of credibility, where people vouch for others' competence on a subject. Not just for licensed/protected topics but for just abou

I honestly hope for a framework of credibility, where people vouch for others' competence on a subject. Not just for licensed/protected topics but for just about anything. You know, like those LinkedIn appraisals (yes, they suck but they're just reflection of the society). That ages old "web of trust" idea. Something without any "mandatory" trusted anchors and no censorship. One must be freely able to put any trust links they believe they want to do, as well as decide for themselves if they trust some established institutions or believe everything is a hoax and decide to trust someone else, say a person or organization we may call a charlatan. Basically, anyone should be able to make any calculations on that graph as they see fit, and use those results for themselves. Just like it already happens. It obviously would be incredibly broken and plagued with all sort of issues, but it would reflect the society. And I'm not sure how it can be made viable - no web of trust except for the original one (the society itself) had ever worked. It probably need some modern fads - be it dancing pigs like Facebook did for their graph, or potential profits of cryptocurrencies, I honestly don't know. And I readily recognize it could lead to privacy problems, especially given the current trend of radical ostracization upon any slightest disagreement. But if it somehow works (it probably won't, sadly) it could be the ultimate solution to licensing boards, fake news, propaganda and many other kinds of memetic storms. But surely, society can't remove licensing requirements before there's a viable alternative. Just a random thought.

mcguire 5y ago
How difficult would it be, do you think, to establish a web-of-trust for those advocating the dangers of vaccines? How large would that web of trust be, and how many people would fall prey to it? Bad money drives out good. Bad speech drives out good, too.
- drdaeman 5y ago
  > and how many people would fall prey to it? How exactly? I could be wrong, but I believe it won't make things any worse. I see three scenarios: First, let's say someone is an antivaxer, already subscribed to that belief. If they're reading some posting, the fact that some other antivaxer vouched for that post's author doesn't change much. If someone had not made any opinion yet - they won't have any trust connections on that matter, so if they're reading an anti-vaccine posting, they won't have any impact. Their query to the web of trust would say "no idea, you don't have any trust anchors set on this matter, go figure out whom you trust first". It is only if they have a friend they trust on the subject vaccines, and that friend happens to be an antivaxer, they may have a "your friend vouches for this" result. This is the only scenario I see, where someone could be swayed towards a certain opinion, and I'd be damned if this doesn't already happens without any technologies, completely organically. It's their peer group that sways them, so the technology in question here doesn't make it worse or better, it just accelerates the process. Again, I could be wrong about either or all of those. I readily recognize all of this stems from my beliefs in humanity (in general) as something that can work out its way through any amount of dissenting opinions, finding out the truths. I mean, historically, society had worked this way. And a controversial belief that accelerating this is a good idea - I admit, I don't have a rational arguments for or against this part - it's just a belief. Guess it comes from hatred of seeing all those slow simmering almost-pertpetual memetic wars and being unhappy about their fallout, yet not believing any forcible silencing of opposing forces is a viable approach. Oh, and just to clarify: it's extremely important that trust must not be ultimate (exceptions apply for closest peers, e.g. partners or parents), but limited to a subject. E.g. I may trust someone on subject X but at the same time know they're avid subscribers to what I think is a false belief on subject Y, so I explicitly distrust them on that matter.
  - mcguire 5y ago
    "If someone had not made any opinion yet - they won't have any trust connections on that matter, so if they're reading an anti-vaccine posting,..." They see that it is vouched for by 1000 other people. They see that a countering article that is also vouched for by 1000 people, but those people are all so-called "experts" and they've read HN and know that "experts" are biased and have perverse incentives and such. Or they just flip a coin. Either way, they believe the anti-vax article. Because the feedback loop is very open, they see no consequences for their choice. So they also vouch for the anti-vax article.
    - drdaeman 5y ago
      > They see that it is vouched for by 1000 other people. Oh, no, no, if I understood your idea - that's something I really want to avoid. This sort of query must explicitly be against the recommended/designed use. It cannot be a popular vote, it must be peer-to-peer. A web rather than a cesspool. A query must result in "no paths to any trusted sources found". Just like in the real world - my opinions on, say, some processes in metallurgy (something I have no clue about) cannot depend on opinion of someone I don't have any connections with, even if they're renown expert in that field. I may trust them through some institutions, certification programs or something like that - say, they graduated from some university that I consider trustworthy. But if there are no connections at all I can't say anything about that person's credibility. Best I can do upon reading their opinion is that it's self-appraised. This web must be designed in a way that people they don't know, people that aren't connected to one's node in the web through peers they know and trust, mustn't have any significant weight, no matter how many there are. One should be free to make their own arbitrary queries, but this online popular vote thing is already known to be a very flawed metric. Not to mention it would require to solve that very hard one-true-identity problem. If all your connections are people you know, and their connections are people they know, identity becomes a non-issue. As long as the small world hypothesis is close enough to the reality, this should work. And it would be a non-issue even if someone invents a fake identity - it'll either a recognized legit alter ego (pseudonymous account, who still may express opinions people listen to), or a meaningless fake person that no one knows (that may vouch for whatever, the only connections is through its creator).
- drdaeman 5y ago
  > Bad speech drives out good, too. Ah, right. I think I got it now. That's where the issue is. It could be that this web of trust just cannot work unless it covers most of the world. If it would, I believe it would be just a technical aid, a machine-coded representation of an already established "organic" web of trust. But it must start from zero, and it cannot grow as it would be lethally poisoned, becoming a place composed mostly of controversial vouches. Could be very true. I need to think about this. One immediate thought is that even a graph of only the "bad voices" (for arbitrary definition of bad) can be perfectly useful, as a sort of noise ("fake news", "lunacy", "bullshit" - however we want to call it) filter. It's not that a web of trust spreads opinions, it spreads relations to their opinions. And a signal that someone vouches directly or indirectly (although "vouching" is not a good term for this case) can be perceived arbitrarily. So maybe it's not that bad. It may be bad by reinforcing echo chambers, but it could also be good in pointing out the connections, those bridges between the factions where arguing is not moot.

mcguire 5y ago

How difficult would it be, do you think, to establish a web-of-trust for those advocating the dangers of vaccines? How large would that web of trust be, and how many people would fall prey to it? Bad money drives out good. Bad speech drives out good, too.

drdaeman 5y ago
> and how many people would fall prey to it? How exactly? I could be wrong, but I believe it won't make things any worse. I see three scenarios: First, let's say someone is an antivaxer, already subscribed to that belief. If they're reading some posting, the fact that some other antivaxer vouched for that post's author doesn't change much. If someone had not made any opinion yet - they won't have any trust connections on that matter, so if they're reading an anti-vaccine posting, they won't have any impact. Their query to the web of trust would say "no idea, you don't have any trust anchors set on this matter, go figure out whom you trust first". It is only if they have a friend they trust on the subject vaccines, and that friend happens to be an antivaxer, they may have a "your friend vouches for this" result. This is the only scenario I see, where someone could be swayed towards a certain opinion, and I'd be damned if this doesn't already happens without any technologies, completely organically. It's their peer group that sways them, so the technology in question here doesn't make it worse or better, it just accelerates the process. Again, I could be wrong about either or all of those. I readily recognize all of this stems from my beliefs in humanity (in general) as something that can work out its way through any amount of dissenting opinions, finding out the truths. I mean, historically, society had worked this way. And a controversial belief that accelerating this is a good idea - I admit, I don't have a rational arguments for or against this part - it's just a belief. Guess it comes from hatred of seeing all those slow simmering almost-pertpetual memetic wars and being unhappy about their fallout, yet not believing any forcible silencing of opposing forces is a viable approach. Oh, and just to clarify: it's extremely important that trust must not be ultimate (exceptions apply for closest peers, e.g. partners or parents), but limited to a subject. E.g. I may trust someone on subject X but at the same time know they're avid subscribers to what I think is a false belief on subject Y, so I explicitly distrust them on that matter.
- mcguire 5y ago
  "If someone had not made any opinion yet - they won't have any trust connections on that matter, so if they're reading an anti-vaccine posting,..." They see that it is vouched for by 1000 other people. They see that a countering article that is also vouched for by 1000 people, but those people are all so-called "experts" and they've read HN and know that "experts" are biased and have perverse incentives and such. Or they just flip a coin. Either way, they believe the anti-vax article. Because the feedback loop is very open, they see no consequences for their choice. So they also vouch for the anti-vax article.
  - drdaeman 5y ago
    > They see that it is vouched for by 1000 other people. Oh, no, no, if I understood your idea - that's something I really want to avoid. This sort of query must explicitly be against the recommended/designed use. It cannot be a popular vote, it must be peer-to-peer. A web rather than a cesspool. A query must result in "no paths to any trusted sources found". Just like in the real world - my opinions on, say, some processes in metallurgy (something I have no clue about) cannot depend on opinion of someone I don't have any connections with, even if they're renown expert in that field. I may trust them through some institutions, certification programs or something like that - say, they graduated from some university that I consider trustworthy. But if there are no connections at all I can't say anything about that person's credibility. Best I can do upon reading their opinion is that it's self-appraised. This web must be designed in a way that people they don't know, people that aren't connected to one's node in the web through peers they know and trust, mustn't have any significant weight, no matter how many there are. One should be free to make their own arbitrary queries, but this online popular vote thing is already known to be a very flawed metric. Not to mention it would require to solve that very hard one-true-identity problem. If all your connections are people you know, and their connections are people they know, identity becomes a non-issue. As long as the small world hypothesis is close enough to the reality, this should work. And it would be a non-issue even if someone invents a fake identity - it'll either a recognized legit alter ego (pseudonymous account, who still may express opinions people listen to), or a meaningless fake person that no one knows (that may vouch for whatever, the only connections is through its creator).
drdaeman 5y ago
> Bad speech drives out good, too. Ah, right. I think I got it now. That's where the issue is. It could be that this web of trust just cannot work unless it covers most of the world. If it would, I believe it would be just a technical aid, a machine-coded representation of an already established "organic" web of trust. But it must start from zero, and it cannot grow as it would be lethally poisoned, becoming a place composed mostly of controversial vouches. Could be very true. I need to think about this. One immediate thought is that even a graph of only the "bad voices" (for arbitrary definition of bad) can be perfectly useful, as a sort of noise ("fake news", "lunacy", "bullshit" - however we want to call it) filter. It's not that a web of trust spreads opinions, it spreads relations to their opinions. And a signal that someone vouches directly or indirectly (although "vouching" is not a good term for this case) can be perceived arbitrarily. So maybe it's not that bad. It may be bad by reinforcing echo chambers, but it could also be good in pointing out the connections, those bridges between the factions where arguing is not moot.