4 ms·

> It is possible to ban one kind of content Actually, have you done that? If you haven't, you should try because in general it's an extremely difficult proble

> It is possible to ban one kind of content Actually, have you done that? If you haven't, you should try because in general it's an extremely difficult problem. The fact that essentially every single platform that has user-generated/-contributed content struggles with this should make it obvious that it's not easy at all. Even the stuff you'd think would be simple (e.g. blocking profanity) is hard, but once you try to nail down things as nebulous as subject matter, it's more or less impossible to define a policy that always works. The most reliable is to have a human apply a "I know it when I see it" standard, but that obviously doesn't scale and is itself subjective. So, if you decide to ban Nazi ideology, that seems good at first. But can someone write a historical post that outlines the ideology so we can learn about it? Seems reasonable. But what if the mere tone of the historical article seems to shine a somewhat positive light on it? What if the article says the ideology is bad overall but had some ideas that were maybe good at their core but then became distorted? It goes on and on and on - for any policy you define, there will be ways for bad stuff to slip through and there will be cases where good stuff gets censored. You may adopt the position that that's good enough, but it is quite hard to "ban one kind of content" effectively.

krapp 3y ago
You're confusing "effectively" with "perfectly." Yes I believe it's possible to effectively censor Nazi ideology without having an objective and universally accepted definition of that, which unambiguously covers all possible edge cases, and despite the possibility of error. We accept a system of laws despite laws being ambiguous, needing humans to recognize context and nuance to interpret. We communicate with language despite words not being defined with the rigidity of mathematical proof. We accept scientific models which fail to perfectly map the territory. To claim that efforts which are imperfect must be abandoned is absurdly defeatist, simply giving in to chaos.
- dbrueck 3y ago
  > You're confusing "effectively" with "perfectly." We'll see! :) Can you please share what you think Substack's content policy should be? Just your rough draft of content guidelines that would "effectively censor Nazi ideology". I'm not talking about vague platitudes but actual rules or guidelines for authors and commenters to follow.
  - krapp 3y ago
    No, I'm not going to play that game. The validity of my argument does not rest on my ability to draft a specific content policy for a website. I'm not an expert in that field and I assume the process of doing so to standard would require far more time and effort than this conversation is worth. Substack's actual content policy is here[0], and as you can read it already contains prohibitions of various kinds of content. Feel free to substitute "Nazi ideology" with "material that advocates, threatens, or shows you causing harm to yourself, other people, or animals" or "sexually exploitative content" if you like. We're just arguing about where to draw the lines in the sand, but clearly drawing those lines is possible. [0]https://substack.com/content https://substack.com/content
    - dbrueck 3y ago
      This whole story is about the fact that the articles you want censored are allowed under Substack's content policy, so I don't think Substack's content policy is a good example for what you're after, no? Like I pointed out at the very beginning, efforts to censor always seem to break down, and one reason for that is that it's hard to come up with coherent policies. It's easy to speak at a super high level about how things ought to be, but once you try to distill those ideas into something that has any practical value - i.e. something that can be actually used and applied - it all starts to fall apart. The result is that platform providers end up in one of 2 main camps: 1) The content does not represent us at all. We set some very broad limits, but for the most part, people can say what they say and we're just providing the platform. 2) Content has to conform to our standards, and application of those standards can and will appear to be completely arbitrary. Neither approach is without its problems, and both approaches anger a lot of people. In general, though, if you're in the business of trying to attract creators and help them build audiences on your platform, #1 is much less bad than #2 - the latter will either limit your ability to get off the ground or will become a millstone when you do.

krapp 3y ago

You're confusing "effectively" with "perfectly." Yes I believe it's possible to effectively censor Nazi ideology without having an objective and universally accepted definition of that, which unambiguously covers all possible edge cases, and despite the possibility of error. We accept a system of laws despite laws being ambiguous, needing humans to recognize context and nuance to interpret. We communicate with language despite words not being defined with the rigidity of mathematical proof. We accept scientific models which fail to perfectly map the territory. To claim that efforts which are imperfect must be abandoned is absurdly defeatist, simply giving in to chaos.

dbrueck 3y ago
> You're confusing "effectively" with "perfectly." We'll see! :) Can you please share what you think Substack's content policy should be? Just your rough draft of content guidelines that would "effectively censor Nazi ideology". I'm not talking about vague platitudes but actual rules or guidelines for authors and commenters to follow.
- krapp 3y ago
  No, I'm not going to play that game. The validity of my argument does not rest on my ability to draft a specific content policy for a website. I'm not an expert in that field and I assume the process of doing so to standard would require far more time and effort than this conversation is worth. Substack's actual content policy is here[0], and as you can read it already contains prohibitions of various kinds of content. Feel free to substitute "Nazi ideology" with "material that advocates, threatens, or shows you causing harm to yourself, other people, or animals" or "sexually exploitative content" if you like. We're just arguing about where to draw the lines in the sand, but clearly drawing those lines is possible. [0]https://substack.com/content https://substack.com/content
  - dbrueck 3y ago
    This whole story is about the fact that the articles you want censored are allowed under Substack's content policy, so I don't think Substack's content policy is a good example for what you're after, no? Like I pointed out at the very beginning, efforts to censor always seem to break down, and one reason for that is that it's hard to come up with coherent policies. It's easy to speak at a super high level about how things ought to be, but once you try to distill those ideas into something that has any practical value - i.e. something that can be actually used and applied - it all starts to fall apart. The result is that platform providers end up in one of 2 main camps: 1) The content does not represent us at all. We set some very broad limits, but for the most part, people can say what they say and we're just providing the platform. 2) Content has to conform to our standards, and application of those standards can and will appear to be completely arbitrary. Neither approach is without its problems, and both approaches anger a lot of people. In general, though, if you're in the business of trying to attract creators and help them build audiences on your platform, #1 is much less bad than #2 - the latter will either limit your ability to get off the ground or will become a millstone when you do.