pull down to refresh
sum(log10()) was great but it relied on counting people, because it favors sybils, and we counted people via trust, which is not perfect either.
I think we all want something democracy-like. As media/news consumers, we want to know what consensus is. Whether we realize it or not, we want algorithms to count people. On the internet, absent KYC, there are few if any objective ways to do that.
Bitcoin's consensus relies on bottom-up economic activity to achieve consensus but, unlike SN, economic activity is the point of bitcoin.
but it relied on counting people, because it favors sybils
I remember - we discussed it when you shipped no-trust and it makes sense that in lieu of anti-sybil mechanics there is no such dampening possible on rank.
As media/news consumers, we want to know what consensus is.
Yes, and this is why it's worth talking about this. I still agree with reputation being a potentially useful tool in this, just it is hard to meet both that and your ELI5 requirement for post ranking, and then to make it non-hierarchical, avoid bubbles / winner-takes-it-all schemes.
I'm still running my mute experiment. I prolonged it because it's bliss. I have not muted any new bots though, and I'm considering removing the existing ones from my list. That wasn't what made this bliss. What did was muting "top" stackers that waste my time. Probably I am wasting yours, so feel free to mute me if I am bothering you. Or just say "zip it opti" - I can handle it.
I still agree with reputation being a potentially useful tool in this
Reputation, as we've defined it, seems to differ from trust in that it isn't subjective. But not being subjective, and with SN being retardedly transparent, that makes it more gameable than trust. So in the ideal of either reputation or WoT we have a dilemma: gameability vs subjectivity.
Or just say "zip it opti" - I can handle it.
Nah! I've found our chats informative. You hold ideals and pragmatics in superposition (as do many stackers) while also understanding uniquely digital problems well.
This is a nice one, but I think we cannot make a subjective thing objective by democracy, that is exactly what makes me shudder in life for a long time now - we're still a pre-truth civilization as we have not been able to consistently separate fact from fiction in day-to-day life. This is why we haven't reached the information age; we're still tribal, territorial primates with a long evolution ahead of us.
What you can do is measure consensus. But that is also not what is happening on SN now, as "money is the moderator", not "stackers' collective opinion is the moderator". The distinction being that SN is open loop, so there is no limit to moderation power other than Bitcoin's built-in limits (assuming that there is no paper bitcoin, and that assumption is likely faulty.)
In my personal utopian scenario: in a decaying position in global
new, a slower decaying position in territorynew, and since no one is interested in it, the bottom of all ranked views. If everyone muted the author, then it will never rise to prominence because it won't get zaps.But if you were asking: how much of a place does it have right now? Then the answer is: the place the author is willing to buy. And this is the issue here. We don't want people manipulating their own ranks with sockpuppets, so instead we gave everyone a button to manipulate their rank. Let me apply that to your metaphor:
The root cause of this problem isn't the mute. The root cause is the boost (or sockpuppet upzapping in lieu of boost - same outcome) and the root enabler of that is open loop money being the unbounded moderator. The muting is just a personal defense mechanism, and I think that undue weight is given to the usage of it. I counted less than 10 stackers (maybe even less than 5) that were actively trying to selflessly counter the big downzap war earlier this year. If it doesn't happen organically, it won't happen as a forced mechanism. If forced, the established base may leave, which is even worse than the new stacker leaving, as the established base provides the content. This will only make it harder to fight "rage bait war slop".
I wish I were better at math, so that I could define a formula that shows the impact of unbounded open-loop. The soothing effect of
sum(log10())ranking we had before no-trust is truly missed, but I suspect that it simply masked the problem by flattening the curve, instead of solving the underlying problem.