Eh, it seems somewhat self-evident that it does not make a lot of sense to expect agents to avoid punishments which do not exist at the time, as such. CEV, to the point that CEV(mankind) wouldn’t be an empty-set anyways, would probably include it just by fiat of it being one of the pillars of the rule of law, and since it’s about “the people we’d want to be”, presumably your CEV at least would contain it, as would mine. There is no relation to TDT, since we’re talking about preferences of groups of agents, not general instrumental rationality.
There is no relation to TDT, since we’re talking about preferences of groups of agents
You mean we aren’t talking about the choice whether or not to punish someone? I don’t see how that holds. If you only discuss a decision theory in the abstract but are not willing to use it for practical decisions than you are likely going to have a bad decision theory.
Don’t compartmentalize and stop using your decision theory when things get political.
In this case punishing people for doing something that’s bad for the community can discourage other people from doing something bad for the community against which we don’t have explicit rules.
Even if I agree that nation state should only punish in a court of law based on explicit rules that doesn’t mean that I think the same is true for privately owned online communities. If I throw a party and someone misbehaves I can throw that person out even without him violating a previously explicitly stated rule. A lot of social interaction works about people simply observing implicit rules and focusing on being nice to each other.
You mean we aren’t talking about the choice whether or not to punish someone?
TDT can’t tell you how to optimally arrange the flavors of an ice cream cone if you don’t input which flavors you like. “But how can that be, that is a choice too?” Decision theory tells you which decisions are optimal, given your preferences. My preference is rule of law (which also makes sense as an instrumental value), I suspect yours is too. TDT doesn’t care. It can’t tell you your preferences (though it can tell you which instrumental values would make sense).
Don’t compartmentalize and stop using your decision theory when things get political.
I don’t understand. TDT isn’t my practical decision theory (I’m a meatbag, not an abstract agent), nor did I bring it up. Nor is it applicable anyways. Optimality is viewpoint dependent.
My preference is rule of law (which also makes sense as an instrumental value), I suspect yours is too.
For nation states with a monopoly on power I consider the rule of law to be valuable but I don’t consider it to be a terminal value for online communities or when I host a party.
In most social interaction punishing people for violating implicit community norms is quite common.
The person who engages in the block downvoting might even think of themselves as punishing someone else for doing something bad.
I don’t understand. TDT isn’t my practical decision theory (I’m a meatbag, not an abstract agent), nor did I bring it up. Nor is it applicable anyways.
If it isn’t applicable then what’s wrong with TDT? How do we fix it?
Even if you don’t personally follow TDT, you are here on LW and while you are here making the argument that the policies you are advocated make sense under TDT has merits if you want to convince others.
Let’s stop with the reference class tennis. This community does have established and explicit rules, such as “no proposing violence, not even hypothetically”. It is not like one of your parties, I suspect. And while you may tell someone to leave you alone, or to get out, I wouldn’t say that official punishments against breaking inofficial “norms” are the rule. At least hopefully nowhere I’d like to be. Note how this community has grappled time and again with coming up with a clearly defined norm on this, which would decohere the congruence even if LW were like a party. Meet-ups have resorted to clip-on notes whether hugging is ok with that person. So much for implicit norms for social interaction.
Someone who engages in block downvoting would be sending a signal, using tools as were provided. What’s the obsession with the punishment-concept? (Warning, flippant aside: Do we need a good public beating, or what?)
If it isn’t applicable then what’s wrong with TDT?
In a word, nothing. If you use TDT going off of “I don’t want agents to be punished for actions against which there are no rules”, then TDT will include that when giving you the optimal course of action. If you use TDT going off of “I don’t care whether agents are punished for actions against which there are no rules”, then it won’t include that. It’s the reason why paperclippers and anti-paperclippers both can use TDT. TDT doesn’t judge your preferences :-).
This community does have established and explicit rules, such as “no proposing violence, not even hypothetically”.
Those rules are not in a place where a new member would easily find them. Some people even think there a rule against politics when there’s no such thing on LW.
Someone who engages in block downvoting would be sending a signal, using tools as were provided. What’s the obsession with the punishment-concept?
You were the person who started speaking about punishment. For my part when I was forum moderator I didn’t think of myself punishing weeds when I tried to rip them out of my healthy garden.
I did ban people but not to punish them but because I thought the forum would be healthier without them.
At least read the explanation of that rule first, would you? There you go:
If we determine that you’re e.g. following a particular user around and leaving insulting comments to them, we reserve the right to delete those comments.
Your leading OP title, including the phrase “mass-downvote harassment” is insincere reasoning, because it is circular. It has never been established whether mass-downvoting should always be considered “harassment”. You’d consider it so. I don’t. Come now, be so courteous as to assume other people have reasons for their behavior.
Not even the wiki, which does include an example, makes mention of mass downvoting even though the topic has come up many times. The reason for that is not “well, we can’t list everything, we don’t list hacking a server, for example”. That would be a ridiculous argument. One is using established feedback mechanisms, one isn’t. New rule: You must always give reasons for each and every vote, otherwise you’ll be publicly shamed for harassment.
Downvotes are a user’s individual and private choice. He/She can use it to confer whatever message he/she so chooses. Don’t like it? Make a rule against it. Such as an upper bound on allowed downvotes. Oh wait, such an upper-bound has already been implemented? And it doesn’t disallow downvoting most of a user’s comments? Maybe your moral intuitions on the matter aren’t as general as you’d like them to be.
Signing off on the topic, though I’ll leave you the last word, if you so choose.
Eh, it seems somewhat self-evident that it does not make a lot of sense to expect agents to avoid punishments which do not exist at the time, as such. CEV, to the point that CEV(mankind) wouldn’t be an empty-set anyways, would probably include it just by fiat of it being one of the pillars of the rule of law, and since it’s about “the people we’d want to be”, presumably your CEV at least would contain it, as would mine. There is no relation to TDT, since we’re talking about preferences of groups of agents, not general instrumental rationality.
You mean we aren’t talking about the choice whether or not to punish someone? I don’t see how that holds. If you only discuss a decision theory in the abstract but are not willing to use it for practical decisions than you are likely going to have a bad decision theory.
Don’t compartmentalize and stop using your decision theory when things get political.
In this case punishing people for doing something that’s bad for the community can discourage other people from doing something bad for the community against which we don’t have explicit rules.
Even if I agree that nation state should only punish in a court of law based on explicit rules that doesn’t mean that I think the same is true for privately owned online communities. If I throw a party and someone misbehaves I can throw that person out even without him violating a previously explicitly stated rule. A lot of social interaction works about people simply observing implicit rules and focusing on being nice to each other.
TDT can’t tell you how to optimally arrange the flavors of an ice cream cone if you don’t input which flavors you like. “But how can that be, that is a choice too?” Decision theory tells you which decisions are optimal, given your preferences. My preference is rule of law (which also makes sense as an instrumental value), I suspect yours is too. TDT doesn’t care. It can’t tell you your preferences (though it can tell you which instrumental values would make sense).
I don’t understand. TDT isn’t my practical decision theory (I’m a meatbag, not an abstract agent), nor did I bring it up. Nor is it applicable anyways. Optimality is viewpoint dependent.
For nation states with a monopoly on power I consider the rule of law to be valuable but I don’t consider it to be a terminal value for online communities or when I host a party. In most social interaction punishing people for violating implicit community norms is quite common.
The person who engages in the block downvoting might even think of themselves as punishing someone else for doing something bad.
If it isn’t applicable then what’s wrong with TDT? How do we fix it?
Even if you don’t personally follow TDT, you are here on LW and while you are here making the argument that the policies you are advocated make sense under TDT has merits if you want to convince others.
Let’s stop with the reference class tennis. This community does have established and explicit rules, such as “no proposing violence, not even hypothetically”. It is not like one of your parties, I suspect. And while you may tell someone to leave you alone, or to get out, I wouldn’t say that official punishments against breaking inofficial “norms” are the rule. At least hopefully nowhere I’d like to be. Note how this community has grappled time and again with coming up with a clearly defined norm on this, which would decohere the congruence even if LW were like a party. Meet-ups have resorted to clip-on notes whether hugging is ok with that person. So much for implicit norms for social interaction.
Someone who engages in block downvoting would be sending a signal, using tools as were provided. What’s the obsession with the punishment-concept? (Warning, flippant aside: Do we need a good public beating, or what?)
In a word, nothing. If you use TDT going off of “I don’t want agents to be punished for actions against which there are no rules”, then TDT will include that when giving you the optimal course of action. If you use TDT going off of “I don’t care whether agents are punished for actions against which there are no rules”, then it won’t include that. It’s the reason why paperclippers and anti-paperclippers both can use TDT. TDT doesn’t judge your preferences :-).
Those rules are not in a place where a new member would easily find them. Some people even think there a rule against politics when there’s no such thing on LW.
You were the person who started speaking about punishment. For my part when I was forum moderator I didn’t think of myself punishing weeds when I tried to rip them out of my healthy garden. I did ban people but not to punish them but because I thought the forum would be healthier without them.
That doesn’t tell you everything there to know about hugging. There are still issues like the length of the hug that isn’t fixed by the rules.
Especially in a community of munchkins you don’t want to allow people to game the rules by moving exactly within them but violating their spirit.
The rules also explicitly include a no harassment of individual users clause.
At least read the explanation of that rule first, would you? There you go:
Your leading OP title, including the phrase “mass-downvote harassment” is insincere reasoning, because it is circular. It has never been established whether mass-downvoting should always be considered “harassment”. You’d consider it so. I don’t. Come now, be so courteous as to assume other people have reasons for their behavior.
Not even the wiki, which does include an example, makes mention of mass downvoting even though the topic has come up many times. The reason for that is not “well, we can’t list everything, we don’t list hacking a server, for example”. That would be a ridiculous argument. One is using established feedback mechanisms, one isn’t. New rule: You must always give reasons for each and every vote, otherwise you’ll be publicly shamed for harassment.
Downvotes are a user’s individual and private choice. He/She can use it to confer whatever message he/she so chooses. Don’t like it? Make a rule against it. Such as an upper bound on allowed downvotes. Oh wait, such an upper-bound has already been implemented? And it doesn’t disallow downvoting most of a user’s comments? Maybe your moral intuitions on the matter aren’t as general as you’d like them to be.
Signing off on the topic, though I’ll leave you the last word, if you so choose.