eugman comments on Anti-akrasia remote monitoring experiment - Less Wrong

45 Post author: cousin_it 27 September 2010 11:34PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (114)

You are viewing a single comment's thread. Show more comments above.

Comment author: AdeleneDawner 29 September 2010 02:58:12PM 8 points [-]

This would need some impressive anti-troll precautions. Or at least a way to temporarily disable the horn, so nobody can blast it when you're in the middle of a video chat with a client, but then that'd be a source of temptation.

Comment author: eugman 29 September 2010 06:59:52PM 2 points [-]

One possible solution is to build a set of training data, which with a tagged series of sceenshots and fake responses. So the watcher wouldn't know for sure whether they were getting a real person or old data. heck, you could use that element to moderate previous watchers.