I’d be interested to know whether LW readers, or adherents of our particular genus of self-optimization scheme more generally, had a different coefficient on that hyperbola than average before being exposed to LW-like ideas. Anecdotal evidence suggests that we’d discount less aggressively, but the sample’s too small and unrepresentative for me to draw any reliable conclusions.
Pessimistically, I would expect exposure to LessWrong would not alter the coefficient of our entire reward scheme, but rather patch certain graphs, as it were
I’d be interested to know whether LW readers, or adherents of our particular genus of self-optimization scheme more generally, had a different coefficient on that hyperbola than average before being exposed to LW-like ideas. Anecdotal evidence suggests that we’d discount less aggressively, but the sample’s too small and unrepresentative for me to draw any reliable conclusions.
Pessimistically, I would expect exposure to LessWrong would not alter the coefficient of our entire reward scheme, but rather patch certain graphs, as it were