After I wrote the first ending, I realized that there was another possibility, so I wrote a second ending as well. The Overcoming Bias readership will be given a chance to determine which one becomes the True Ending and which the alternative—though it won’t be as easy as a vote.
… and anesthetized the Lord Pilot.
… [This option will become the True Ending only if someone suggests it in the comments before the previous ending is posted tomorrow. Otherwise, the first ending is the True one.]
Insofar as definitions can be right or wrong, so also counterfactual consequences can be right or wrong, and thus fictional evidence can be right or wrong…
Well, I’m glad the story wasn’t ruined by the alternative being too obvious. If no one’s thought of it yet in the comments, then it’s at least plausible that the people on the ship didn’t think of it earlier.
So the rightnesses of the two bodies of fictional evidence in the two endings both depend on the audience’s skill at applied metaethics? And you want to increase the expected rightness of the true ending by correlating the true ending with the audience’s unknown skill? Or by giving the audience an incentive to increase their skill?
(I don’t know the solution. This comment reasoning about your motives is to narrow the search space. Plus it proposes a meaning for your otherwise unexplained term “True”.)
I don’t understand. If it is not known which model is correct, can’t a Bayesian choose policies by the predictive distributions of consequences after marginalizing out the choice of model? Robin seems to be invoking an academic norm of only using vetted quantitative models on important questions, and he seems to be partly expecting that the intuitive force of this norm should somehow result in an agreement that his position is epistemically superior. Can’t the intuitive force of the norm be translated into a justification in something like the game theory of human rhetoric? For example, perhaps the norm is popular in academia because everyone half-consciously understands that the norm is meant to stop people from using the strategy of selecting models which lead to emotionally compelling predictions? Is there a more optimal way to approximate the contributions (compelling or otherwise) of non-vetted models to an ideal posterior belief? If Eliezer is breaking a normal procedural safeguard in human rhetoric, one should clarify the specific epistemic consequences that should be expected when people break that safeguard, and not just repeatedly point out that he is breaking it.