8 ms·

I think you're spot on regarding the free/open-access platform inviting this kind of thing. I don't know if there really is a feasible solution - maybe a delaye

I think you're spot on regarding the free/open-access platform inviting this kind of thing. I don't know if there really is a feasible solution - maybe a delayed-broadcast setup where a human moderator reviews each feed/view? Also maybe automatically have a human attach and watch feeds that are spontaneously getting popular without a clear reason, etc. I've read some folks floating the thought of requiring a Gov ID to livestream, etc. I am not in favor of this. How would that stop a psycho like this? He wouldn't care if you knew his real id. He would just pretend to be docile / normal until the second he starts broadcasting his horrible intentions.

CharlesColeman 8y ago
> I don't know if there really is a feasible solution - maybe a delayed-broadcast setup where a human moderator reviews each feed/view? Also maybe automatically have a human attach and watch feeds that are spontaneously getting popular without a clear reason, etc. I think those are very good ideas for mitigating this kind of problem. Facebook (and other internet companies) have basically built sharing platforms with no breaks, and that opens them up to exploitation by people like this mass shooter. I think the only effective solutions will involve retrofitting breaks and proactively inserting human moderators into the loop. There's literally no good reason that something should be able to go viral globally within hours of creation, not even disconnected natural disaster reports.
- ng12 8y ago
  > There's literally no good reason that something should be able to go viral globally within hours of creation, not even disconnected natural disaster reports. The article states the stream had ~200 views total. That's not anywhere close to viral.
  - CharlesColeman 8y ago
    > The article states the stream had ~200 views total. That's not anywhere close to viral. No. According to the article the original video actually had ~4,000 views in the unspecified amount of time before it was removed. Only ~200 people watched it live. But you missed my point: the current design is setup to present as few obstacles as possible to something going viral, which means it's easy for something to spread so quickly it's impossible to moderate. The solution is to add obstacles on the path to virality to prevent things from going viral. These could include things like post-visibility delays that could trigger even before the objectionable content is even close to viral (and would similarly slow down the spread of mutated reposts).
    - deno 8y ago
      The video went “viral” because of news coverage. This idea that no one would have seen this video without Facebook is laughable. The only real “solutions” are the ones presented in parent post, that is total censorship beyond what even China’s doing. The scary thing is about half the people in this thread would actually welcome it. I really don’t worry about mass shootings, because statistically they’re still utterly insignificant. You can take a step back and see the sky’s not (objectively) falling. But every time something like this happens and people race to give up hard-won rights and invite, even demand more control and censorship, it’s hard not to be cynical…
    - ng12 8y ago
      So you want to slow the spread of things that people might be interested because the things they are interested in might be bad? Fair enough re: 4,000 views, but I still think it's reasonable proof that the system works. It's not like millions had seen the video before it got purged.
      - CharlesColeman 8y ago
        > So you want to slow the spread of things that people might be interested because the things they are interested in might be bad? Yes, because for good things the present speed is not necessary and for bad things it makes needed moderation more difficult. My guess is the only people who'd even notice the slowdown would be social media junkies (as in addicted, compulsive over-users).
save_ferris 8y ago
I agree with you, both on your ideas and your skepticism of ID-driven live-streams. One major problem is that companies like FB are financially disincentived to do preemptive content moderation because it would dramatically reduce the rate at which content is generated, and thus the amount of content generated. Maintaining a high rate of content generation keeps engagement high, which in turn drives ad revenue.

CharlesColeman 8y ago

> I don't know if there really is a feasible solution - maybe a delayed-broadcast setup where a human moderator reviews each feed/view? Also maybe automatically have a human attach and watch feeds that are spontaneously getting popular without a clear reason, etc. I think those are very good ideas for mitigating this kind of problem. Facebook (and other internet companies) have basically built sharing platforms with no breaks, and that opens them up to exploitation by people like this mass shooter. I think the only effective solutions will involve retrofitting breaks and proactively inserting human moderators into the loop. There's literally no good reason that something should be able to go viral globally within hours of creation, not even disconnected natural disaster reports.

ng12 8y ago
> There's literally no good reason that something should be able to go viral globally within hours of creation, not even disconnected natural disaster reports. The article states the stream had ~200 views total. That's not anywhere close to viral.
- CharlesColeman 8y ago
  > The article states the stream had ~200 views total. That's not anywhere close to viral. No. According to the article the original video actually had ~4,000 views in the unspecified amount of time before it was removed. Only ~200 people watched it live. But you missed my point: the current design is setup to present as few obstacles as possible to something going viral, which means it's easy for something to spread so quickly it's impossible to moderate. The solution is to add obstacles on the path to virality to prevent things from going viral. These could include things like post-visibility delays that could trigger even before the objectionable content is even close to viral (and would similarly slow down the spread of mutated reposts).
  - deno 8y ago
    The video went “viral” because of news coverage. This idea that no one would have seen this video without Facebook is laughable. The only real “solutions” are the ones presented in parent post, that is total censorship beyond what even China’s doing. The scary thing is about half the people in this thread would actually welcome it. I really don’t worry about mass shootings, because statistically they’re still utterly insignificant. You can take a step back and see the sky’s not (objectively) falling. But every time something like this happens and people race to give up hard-won rights and invite, even demand more control and censorship, it’s hard not to be cynical…
  - ng12 8y ago
    So you want to slow the spread of things that people might be interested because the things they are interested in might be bad? Fair enough re: 4,000 views, but I still think it's reasonable proof that the system works. It's not like millions had seen the video before it got purged.
    - CharlesColeman 8y ago
      > So you want to slow the spread of things that people might be interested because the things they are interested in might be bad? Yes, because for good things the present speed is not necessary and for bad things it makes needed moderation more difficult. My guess is the only people who'd even notice the slowdown would be social media junkies (as in addicted, compulsive over-users).

save_ferris 8y ago

I agree with you, both on your ideas and your skepticism of ID-driven live-streams. One major problem is that companies like FB are financially disincentived to do preemptive content moderation because it would dramatically reduce the rate at which content is generated, and thus the amount of content generated. Maintaining a high rate of content generation keeps engagement high, which in turn drives ad revenue.