= 27d567bc2d48d9f40e9ccc20ea370b15)
Qiaochu_Yuan comments on Open Thread, March 1-15, 2013 - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (237)
Recent experiences have suggested to me that there is a positive correlation between rationality and prosopagnosia. One hypothesis is that dealing with prosopagnosia requires using Bayes to recognize people, so it naturally provides a training ground for Bayesian reasoning. But I'm curious about other possible hypotheses as well as additional anecdotal evidence for or against this conclusion.
What were the recent experiences?
I learned that a surprising number of people involved with CFAR / MIRI have prosopagnosia. (Well, either that or I'm miscalibrated about the prevalence of prosopagnosia.)
How prevalent do you think it is?
I know 4 (I think?) people with prosopagnosia and maybe 800 people total, so my first guess is 0.5%. Wikipedia says 2.5% and the internet says it's difficult to determine the true prevalence because many people don't realize they have it (generalizing from one example, I assume). The observed prevalence in CFAR / MIRI is something like 25%?
So another plausible hypothesis is that rationalists are unusually good at diagnosing their own prosopagnosia and the actual base rate is higher than one would expect based on self-reports.
That is a big difference.
Theory off the top of my head: The causation is in the wrong direction. People who are rational are far more likely to be very systems-oriented, have limited social experiences as children (by having different interests and/or being too dang smart), be highly introverted, and other factors that correlate with being around other people a lot less than your typical person. There's nothing wrong with our hardware per se, it's just that we missed out on critical training data during the learning period,
Anecdotal: I have mild prosopagnosia. I have a lot of trouble recognising people outside their expected context, I make heavy use of non-facial cues. I'm pretty good at putting specific names to specific faces on demand when it feels important enough, although see prev point about expected context. I don't feel like I use anything resembling Bayesian reasoning, I feel like I have the same sense of recognition that I imagine most people have, it's just less dependent on seeing their face and more on other traits (most typically voice and manner of movement).