Keep in mind that the policy is going to be done through human implementation with the specific intention of avoiding inconveniently broad interpretations like this.
It’s not enough to show that the censorship policy could theoretically be used to stifle conversation we actually want here, the important question is whether it actually would be.
It’s not enough to show that the censorship policy could theoretically be used to stifle conversation we actually want here, the important question is whether it actually would be.
I think that’s a very dangerous idea. This community is about developing FAI. FAI should be expected to act according to the rules that you give it. I think the policy should be judged by the way it would actually work if it would be applied as it’s proposed.
There also the problem of encouraging group think: Advocating of illegal conduct gets censored when it goes against the morality of this group but is allowed when it’s within the realms of that morality is bad.
This community should have consistent rules about which discussions are allowed and which aren’t. Censoring on a case by case is problematic.
If you start censoring certain speech that advocates violence and avoid censoring other speech that advocates violence you also have the problem that you get more responsibility for the speech that you allow.
In the absence of a censorship policy you don’t endorse a viewpoint by not censoring the viewpoint. If you however do censor, and you do make decisions not to censor specific speech that’s an endorsement of that speech.
The way it’s proposed is to be applied according to the judgment of a moderator. It makes no sense to pretend that we’re beholden to the strictest letter of the rule when that’s not how it’s actually going to work.
What speech that advocates violence do you think would get a pass while the rest would get censored?
I don’t know excatly how much speech Eliezier wants to censor. I wrote a post with a bunch of examples. I would like to see with of those example Eliezer considers worthy of censorship.
Keep in mind that the policy is going to be done through human implementation with the specific intention of avoiding inconveniently broad interpretations like this.
It’s not enough to show that the censorship policy could theoretically be used to stifle conversation we actually want here, the important question is whether it actually would be.
I think that’s a very dangerous idea. This community is about developing FAI. FAI should be expected to act according to the rules that you give it. I think the policy should be judged by the way it would actually work if it would be applied as it’s proposed.
There also the problem of encouraging group think: Advocating of illegal conduct gets censored when it goes against the morality of this group but is allowed when it’s within the realms of that morality is bad.
This community should have consistent rules about which discussions are allowed and which aren’t. Censoring on a case by case is problematic.
If you start censoring certain speech that advocates violence and avoid censoring other speech that advocates violence you also have the problem that you get more responsibility for the speech that you allow.
In the absence of a censorship policy you don’t endorse a viewpoint by not censoring the viewpoint. If you however do censor, and you do make decisions not to censor specific speech that’s an endorsement of that speech.
The way it’s proposed is to be applied according to the judgment of a moderator. It makes no sense to pretend that we’re beholden to the strictest letter of the rule when that’s not how it’s actually going to work.
What speech that advocates violence do you think would get a pass while the rest would get censored?
I don’t know excatly how much speech Eliezier wants to censor. I wrote a post with a bunch of examples. I would like to see with of those example Eliezer considers worthy of censorship.