A point made by many Bayesians is that it does not matter too much if different people start from widely different priors, since if they update on a large amount of independent evidence, their posteriors will converge. This is often true, though not in all cases, especially if the evidence itself is disputed.
