2 ms·

Don't get me wrong, in no way am I suggesting this is an easy problem to fix. I'm aware of how good people are at cramming huge amounts of knowledge into a silo

Don't get me wrong, in no way am I suggesting this is an easy problem to fix. I'm aware of how good people are at cramming huge amounts of knowledge into a siloed yet consistent narrative (and I am not claiming myself to be completely innocent of this either, though I think I try fairly hard not to). I'm also well aware of the extreme difficulty in making a system that can automatically check for consistency in the way I'm describing. However, there are chinks in ideological armor which, if given the right light, can cause breaches and get people to collaborate and understand one another better. I would argue that the current lack of any form of reliable, extensive consistency checker that people can trust (ie, a SYSTEM they can trust rather than a group of people) leads people to silo themselves into small, isolated networks. They don't operate outside of their network, even though there might be information out there somewhere that's valuable and would cause their worldview to expand, because they know there are at least some nefarious actors putting out inconsistent/untrue information, and it's too hard to identify. If there were some way to automatically "expand" your axioms and see how much of the world they can explain while maintaining their consistency, it could theoretically be a non biased, trustworthy way of testing your assumptions of the world outside of your own, small, limited network, and might help people recognize what parts of the world it is and isn't explaining.

pixl97 7y ago
I'm not sure any human system is self consistent. Our entire society is built on incomplete information.
- blotter_paper 7y ago
  Indeed, I've long desired such a tool not as a weapon to convince my enemies of their logical flaws but as a mirror to convince myself of my own.
- didericis 7y ago
  > I'm not sure any human system is self consistent I think you're right, with the possible exception of math. The entire discipline is about finding a large, self consistent set of rules, and a lot of progress has been made towards that end. I think a system that checks for consistency with regards to observations about the real world would be useful as a roadmap even if the ultimate goal of finding a set of rules that was entirely self consistent and described pretty much everything we observed would be too large/complicated to figure out. If there were a way of testing the effectiveness of our information and our assumptions by testing our information and assumptions against all other information, I think it would help identify how "good" our information and assumptions were. Such a system could theoretically help identify bad actors. If an actor is putting out information that is inconsistent with other information, that could mean a) that actor's information is "bad" b) the other information is "bad". To determine which person is "right", you could look at whether the axioms they put forward are consistent with things you accept as true, how many things those axioms can explain accurately, and then compare that to the axioms and information put forward by other people. I realize this is essentially what people do now when they generate their own little web of trust (although there are lots of charisma/confidence/rhetorical angles to human trust that this is unrelated to), but I think if you could some how automate those consistency checks, you'd be able to more easily test your own ideas without worrying about someone with an opposing ideological bent messing with you/being turned of by personality types of presentation of people with a different worldview/etc. This is all extremely theoretical, and again, if the kind of system I was describing were somehow possible, there's a fair chance it could backfire/be used as a sort of "ultimate truth authority" that ideologues could use to stifle ideas they don't like. But I do think the current situation is such that people are going to try to make a "truth engine" regardless of whether it's possible or a good idea, so trying to figure out how to steer that ship in the best direction possible seems prudent (and imo that direction is to try to make it free/open and dependent on consistency/webs of trust rather than anything authoritative).

pixl97 7y ago

I'm not sure any human system is self consistent. Our entire society is built on incomplete information.

blotter_paper 7y ago
Indeed, I've long desired such a tool not as a weapon to convince my enemies of their logical flaws but as a mirror to convince myself of my own.
didericis 7y ago
> I'm not sure any human system is self consistent I think you're right, with the possible exception of math. The entire discipline is about finding a large, self consistent set of rules, and a lot of progress has been made towards that end. I think a system that checks for consistency with regards to observations about the real world would be useful as a roadmap even if the ultimate goal of finding a set of rules that was entirely self consistent and described pretty much everything we observed would be too large/complicated to figure out. If there were a way of testing the effectiveness of our information and our assumptions by testing our information and assumptions against all other information, I think it would help identify how "good" our information and assumptions were. Such a system could theoretically help identify bad actors. If an actor is putting out information that is inconsistent with other information, that could mean a) that actor's information is "bad" b) the other information is "bad". To determine which person is "right", you could look at whether the axioms they put forward are consistent with things you accept as true, how many things those axioms can explain accurately, and then compare that to the axioms and information put forward by other people. I realize this is essentially what people do now when they generate their own little web of trust (although there are lots of charisma/confidence/rhetorical angles to human trust that this is unrelated to), but I think if you could some how automate those consistency checks, you'd be able to more easily test your own ideas without worrying about someone with an opposing ideological bent messing with you/being turned of by personality types of presentation of people with a different worldview/etc. This is all extremely theoretical, and again, if the kind of system I was describing were somehow possible, there's a fair chance it could backfire/be used as a sort of "ultimate truth authority" that ideologues could use to stifle ideas they don't like. But I do think the current situation is such that people are going to try to make a "truth engine" regardless of whether it's possible or a good idea, so trying to figure out how to steer that ship in the best direction possible seems prudent (and imo that direction is to try to make it free/open and dependent on consistency/webs of trust rather than anything authoritative).