I would be interested in seeing what happens if someone creates a phpBB style AI Safety forum. If it works better than what we have, we should probably just switch the existing stuff we have towards a more similar architecture (though of course, the right tradeoffs might depend on the number of users, so maybe the right choice is to have both architectures in parallel).
Another point is that if LW and a hypothetical phpBB forum have different “cognitive styles”, it could be valuable to keep both around for the sake of cognitive diversity.
I would be interested in seeing what happens if someone creates a phpBB style AI Safety forum. If it works better than what we have, we should probably just switch the existing stuff we have towards a more similar architecture (though of course, the right tradeoffs might depend on the number of users, so maybe the right choice is to have both architectures in parallel).
Another point is that if LW and a hypothetical phpBB forum have different “cognitive styles”, it could be valuable to keep both around for the sake of cognitive diversity.