We literally cannot fit enough possibilities in our head for the truth to reliably be inside our considered hypothesis space
If I understand correctly what it’s referring to, this is not relevantly true; infra-bayesianism solves non-realizability formally, and seems much closer than Bayesian agents to how humans formulate hypotheses (i.e., convex sets of probability distributions over states of the world instead of just probability distributions over states of the world; that allows us to have a hypothesis about some part of the world that we can assign probabilities to independently from other parts of the world).
Wow, this is awesome, thanks for doing it!
If I understand correctly what it’s referring to, this is not relevantly true; infra-bayesianism solves non-realizability formally, and seems much closer than Bayesian agents to how humans formulate hypotheses (i.e., convex sets of probability distributions over states of the world instead of just probability distributions over states of the world; that allows us to have a hypothesis about some part of the world that we can assign probabilities to independently from other parts of the world).