To clarify, I’m not trying to speak out against the perspectives people like Lumifer and VoiceOfRa offer, which I am generally sympathetic to. I think their perspectives are valuable.
Really? Their “perspective” appears to consist in attempting to tear down any hopes, beliefs, or accomplishments someone might have, to the point of occasionally just making a dumb comment out of failure to understand substantive material.
Of course, I stated that a little too disparagingly, but see below...
In general, I agree with the view that Paul Graham has advanced re: Hacker News moderation: on a group rationality level, in an online forum context, civility & niceness end up being very important.
Not just civility and niceness, but affirmative statements. That is, if you’re trying to achieve group epistemic rationality, it is important to come out and say what one actually believes. Statistical learning from a training-set of entirely positive or entirely negative examples is known to be extraordinarily difficult, in fact, nigh impossible (modulo “blah blah Solomonoff”) to do in efficient time.
I think a good group norm is, “Even if you believe something controversial, come out and say it, because only by stating hypotheses and examining evidence can we ever update.” Fully General Critique actually induces a uniform distribution across everything, which means one knows precisely nothing.
Besides which, nobody actually has a uniform distribution built into their real expectations in everyday life. They just adopt that stance when it comes time to talk about Big Issues, because they’ve heard of how Overconfidence Is Bad without having gotten to the part where Systematic Underconfidence Makes Reasoning Nigh-Impossible.
Really? Their “perspective” appears to consist in attempting to tear down any hopes, beliefs, or accomplishments someone might have, to the point of occasionally just making a dumb comment out of failure to understand substantive material.
Of course, I stated that a little too disparagingly, but see below...
Not just civility and niceness, but affirmative statements. That is, if you’re trying to achieve group epistemic rationality, it is important to come out and say what one actually believes. Statistical learning from a training-set of entirely positive or entirely negative examples is known to be extraordinarily difficult, in fact, nigh impossible (modulo “blah blah Solomonoff”) to do in efficient time.
I think a good group norm is, “Even if you believe something controversial, come out and say it, because only by stating hypotheses and examining evidence can we ever update.” Fully General Critique actually induces a uniform distribution across everything, which means one knows precisely nothing.
Besides which, nobody actually has a uniform distribution built into their real expectations in everyday life. They just adopt that stance when it comes time to talk about Big Issues, because they’ve heard of how Overconfidence Is Bad without having gotten to the part where Systematic Underconfidence Makes Reasoning Nigh-Impossible.