4 ms·

Moderation is needed for any communication platform. Flag all keywords you would need to do a bomb threat for review. Preferably have a client side filter, li

Moderation is needed for any communication platform. Flag all keywords you would need to do a bomb threat for review. Preferably have a client side filter, like profanity, so users can customize to their liking and you aren't censoring too heavily. Some server side filtering for illegal messages like threats. Open mod logs for even more transparency.

ceejayoz 5y ago
YikYak found out how immensely difficult that is. So have Facebook and Twitter, despite billions of dollars to throw at such problems. There was a link a few days ago on the HN front page that rather humorously covered the difficulty of this. http://habitatchronicles.com/2007/03/the-untold-history-of-toontowns-speedchat-or-blockchattm-from-disney-finally-arrives/ http://habitatchronicles.com/2007/03/the-untold-history-of-t... "We thought it was the perfect solution, until we set our first 14-year old boy down in front of it. Within minutes he’d created the following sentence: I want to stick my long-necked Giraffe up your fluffy white bunny. "
- hunterb123 5y ago
  Does it really matter if messages like that get through? Client side AI filters can attempt to get rid of raunchy messages. Keyword filters and flagging for things like threats can stop the illegal stuff. Don't try to over censor and nanny state people. edit: to clarify, messages "like that" refer to the kid's message in your above post.
  - ceejayoz 5y ago
    > Does it really matter if messages like that get through? When they're veiled bomb threats against schools, yes. That gets media and regulatory attention that kills the brand - again, as YikYak quickly found out. > Keyword filters and flagging for things like threats can stop the illegal stuff. Again, the idea that keyword filters can handle this is silly. It was silly a decade ago. The big social networks have invested enormous time, money, and effort into increasingly sophisticated AI... and the result is people still get incorrectly and absurdly flagged for innocuous posts while clearly offending ones slip through.

ceejayoz 5y ago

YikYak found out how immensely difficult that is. So have Facebook and Twitter, despite billions of dollars to throw at such problems. There was a link a few days ago on the HN front page that rather humorously covered the difficulty of this. http://habitatchronicles.com/2007/03/the-untold-history-of-toontowns-speedchat-or-blockchattm-from-disney-finally-arrives/ http://habitatchronicles.com/2007/03/the-untold-history-of-t... "We thought it was the perfect solution, until we set our first 14-year old boy down in front of it. Within minutes he’d created the following sentence: I want to stick my long-necked Giraffe up your fluffy white bunny. "

hunterb123 5y ago
Does it really matter if messages like that get through? Client side AI filters can attempt to get rid of raunchy messages. Keyword filters and flagging for things like threats can stop the illegal stuff. Don't try to over censor and nanny state people. edit: to clarify, messages "like that" refer to the kid's message in your above post.
- ceejayoz 5y ago
  > Does it really matter if messages like that get through? When they're veiled bomb threats against schools, yes. That gets media and regulatory attention that kills the brand - again, as YikYak quickly found out. > Keyword filters and flagging for things like threats can stop the illegal stuff. Again, the idea that keyword filters can handle this is silly. It was silly a decade ago. The big social networks have invested enormous time, money, and effort into increasingly sophisticated AI... and the result is people still get incorrectly and absurdly flagged for innocuous posts while clearly offending ones slip through.