“weird” vs “normal”. This concept seems to bundle together “good” and “usual”, or at least “bad” with “unusual”.
If most people are mostly strategic most of the time, then common actions will indeed be strategic ones, so uncommon actions are probably unstrategic. However, in reality, we all have severely bounded rationality (compared to a bayesian superintelligence). “To be human is to make ten thousand errors. No one in this world achieves perfection.” This limits the utility of the absurdity heuristic for judging the utility of actions.
Even if you aren’t making this conflation, “weird” vs “normal” can encourage a map/territory error, where “weird” things are thought of as inherently low-probability, and “normal” things inherently high-probability. Bayesians think of probabilities as a property of observing agents rather than inherent to things-in-themselves. To avoid this mistake, people sometimes say things like “since the beginning, not one unusual thing has ever happened” (which can be interpreted as saying that, if we insist on attaching weirdness/normalness as an inherent property of events, we should consider everything that actually happens as normal).
I think the label “weird” also seems to serve as a curiosity-stopper, in some cases, because the inherent weirdness “explains” the unusual observations. EG, “they’re just weird”.
“weird” vs “normal”. This concept seems to bundle together “good” and “usual”, or at least “bad” with “unusual”.
If most people are mostly strategic most of the time, then common actions will indeed be strategic ones, so uncommon actions are probably unstrategic. However, in reality, we all have severely bounded rationality (compared to a bayesian superintelligence). “To be human is to make ten thousand errors. No one in this world achieves perfection.” This limits the utility of the absurdity heuristic for judging the utility of actions.
Even if you aren’t making this conflation, “weird” vs “normal” can encourage a map/territory error, where “weird” things are thought of as inherently low-probability, and “normal” things inherently high-probability. Bayesians think of probabilities as a property of observing agents rather than inherent to things-in-themselves. To avoid this mistake, people sometimes say things like “since the beginning, not one unusual thing has ever happened” (which can be interpreted as saying that, if we insist on attaching weirdness/normalness as an inherent property of events, we should consider everything that actually happens as normal).
I think the label “weird” also seems to serve as a curiosity-stopper, in some cases, because the inherent weirdness “explains” the unusual observations. EG, “they’re just weird”.