![]() |
Grand Staircase Escalante National Monument |
When used correctly, Bayes' Theorem is great. (Explanation with an example; video discussion.) Once you understand it, it can help you deal with the world more efficiently and accurately.
One simple recent example: Last night, I heard something between a creak and a bang somewhere in the house. Now that could be something bad (a pipe burst, someone breaking in) so should I get up and go explore?
Having lived in this house for 16 years, I know from previous experience that the house is almost always making some noise. In the early years, I would get up and explore; later, I would look around in the morning.
Thus, based on the last 16 years, my "prior" probability is near zero that any sound I hear in the night is something bad. So roll over and keep trying to get to sleep.
(The two links above do a much better job of explaining.)
However, "Bayes" has been discovered by people just smart enough to be dangerous. They point to mistakes (or alleged mistakes) made by "experts," and claim that this means they have to do their own calculations, "using" Bayes.
Of course, this doesn't mean a rigorous, expert analysis of all the data - they've "proven" experts can't be trusted, so the data is skewed! This means anyone can make up any number they want for their "prior," and that number just happens to give them the answer they want. They have mathematically "proven" what Fox News wants to hear.
An otherwise smart but incredibly arrogant economist (there sure are a lot of those) recently did just this. Set his priors high for lab leak, and thus, he has shown that Covid started with a lab leak. Same process to show climate change predictions are wrong, school closures were wrong, and masks don't work. (I really want to see him stick with that when he goes in for surgery. "Take off the mask, Doc! You know: freedom!!"
This is just another example that education doesn't bring people closer to the truth - it just makes us more certain of our political views. Sadly, this has been shown over and over; e.g., Individuals with greater science literacy and education have more polarized beliefs on controversial science topics:
"We find that beliefs are correlated with both political and religious identity for stem cell research, the Big Bang, and human evolution, and with political identity alone on climate change. Individuals with greater education, science education, and science literacy display more polarized beliefs on these issues."
Yet another reason to hope for the rise of robot overlords.
(BTW, I worked with the second author of that paper, back when I was at Carnegie Mellon.)
No comments:
Post a Comment