I agree up to the first half of step 6, but I think the conclusion is wrong (or at least not justified from the argument). There are two different principles involved here:
A finite agent must use an “inductive-ish” prior with a finite complexity
One should use the simplest prior. (Occam’s Razor)
If every finite agent must use an “inductive-ish” prior, then there is no need to invoke or appeal to Occam’s Razor to explain or justify our own inductive tendencies, so Rob’s argument actually undercuts Occam’s Razor.
If we replace Occam’s Razor with the principle that every finite agent must use a prior with finite complexity, then one’s prior is just whatever it is, and not necessarily the simplest prior. There is no longer an argument against someone who says their prior assigns a greater weight to Christianity than to string theory. (In the second half of step 6, Rob says that’s “contrived”, but they could always answer “so what?”)
I agree up to the first half of step 6, but I think the conclusion is wrong (or at least not justified from the argument). There are two different principles involved here:
A finite agent must use an “inductive-ish” prior with a finite complexity
One should use the simplest prior. (Occam’s Razor)
If every finite agent must use an “inductive-ish” prior, then there is no need to invoke or appeal to Occam’s Razor to explain or justify our own inductive tendencies, so Rob’s argument actually undercuts Occam’s Razor.
If we replace Occam’s Razor with the principle that every finite agent must use a prior with finite complexity, then one’s prior is just whatever it is, and not necessarily the simplest prior. There is no longer an argument against someone who says their prior assigns a greater weight to Christianity than to string theory. (In the second half of step 6, Rob says that’s “contrived”, but they could always answer “so what?”)