One think always bothered me about this post. Its main thrust appears to be that one should use (something like) a Komogorov prior rather than one’s intuitive prior because that is what the universe looks like, except we then adjust this prior with the same evidence we used to determine what the prior should be. That seems like double-counting evidence.
One think always bothered me about this post. Its main thrust appears to be that one should use (something like) a Komogorov prior rather than one’s intuitive prior because that is what the universe looks like, except we then adjust this prior with the same evidence we used to determine what the prior should be. That seems like double-counting evidence.