4 ms·

Why not use machine learning and a sort of "reputation system" to filter out the noise and trolls? The system could analyse comments, but the users could also,

by remir 9y ago

Why not use machine learning and a sort of "reputation system" to filter out the noise and trolls? The system could analyse comments, but the users could also, instead of upvote/downvote buttons, have buttons with reactions such as "Interesting", "Insightful", "Not relevant", etc... Some people have informed opinion/interesting things to say and some people just contribute noise to the discussion. The system can learn both from the language and reactions from people. Just like in "real life", if someone is always talking out their ass, you learn to listen to them less and less because they don't have anything meaningful to say. It's just noise. The comment system should do the same. It learns from multiple comments and their reactions to determine a reputation. Obviously, if your comments are always seen as "not-relevant" and "toxic", then your contribution is worth less than other "insightful" and "interesting" contributions, so past a certain threshold, you stop being able to comment.

colejohnson66 9y ago
Isn't that what Facebook does with their curated newsfeed? Doesn't that just make echo chambers?
- remir 9y ago
  I'm not on Facebook so I don't know much about the newsfeed, but the goal is to eliminate trolling and toxicity, not debates.
  - interstitial 9y ago
    The goal is to make companies pay for sponsored post and keep users "engaged" through manipulation. Nothing else.
colordrops 9y ago
Seems dangerous if the rules are not adjusted right. It could end up amplifying certain narratives and extinguishing others, and causing a divide between sets of people as they move to communities where they don't get blocked. Furthermore an AI would be more opaque than a simple voting system, and could easily be manipulated for unethical purposes. Lastly, this sort of system would heavily bias towards the mainstream viewpoint, which is a good heuristic, but the mainstream isn't always right about everything. There is a place for minority voices. I would posit that the extreme political division we now see in the US is a result of more powerful systems for influencing online discourse.
- remir 9y ago
  I get where you're coming from. I guess it depends on how the system is made, but there's a risk, absolutely.

by remir 9y ago

colejohnson66 9y ago

Isn't that what Facebook does with their curated newsfeed? Doesn't that just make echo chambers?

remir 9y ago
I'm not on Facebook so I don't know much about the newsfeed, but the goal is to eliminate trolling and toxicity, not debates.
- interstitial 9y ago
  The goal is to make companies pay for sponsored post and keep users "engaged" through manipulation. Nothing else.

colordrops 9y ago

Seems dangerous if the rules are not adjusted right. It could end up amplifying certain narratives and extinguishing others, and causing a divide between sets of people as they move to communities where they don't get blocked. Furthermore an AI would be more opaque than a simple voting system, and could easily be manipulated for unethical purposes. Lastly, this sort of system would heavily bias towards the mainstream viewpoint, which is a good heuristic, but the mainstream isn't always right about everything. There is a place for minority voices. I would posit that the extreme political division we now see in the US is a result of more powerful systems for influencing online discourse.

remir 9y ago
I get where you're coming from. I guess it depends on how the system is made, but there's a risk, absolutely.