zlacker

I also do not understand what you’re trying to get at with “internalized probabilities” etc. I understand the importance of this sort of jargon to the ‘ratsphere’ and all that (https://www.reddit.com/r/sgiwhistleblowers/s/nLaIGJbWAI), but that doesn’t make it any more intelligible to me. I guess that isn’t the point.

>>aaronb+(OP)
The goal is to update beliefs in all areas when they change in one spot.

As a hypothetical, let's say you believe from prior experience that being mugged has a very high probability. Let's say 50% because it's easier.

Let's also say your friend points out that you've left your home hundreds of times this year and haven't been mugged. 50% seems like a ridiculous overestimate.

Reexamining your priors would involve not only changing your mind about the chance of being mugged, but changing downstream beliefs that might be influenced by that belief (such as what public policies you support).