I’m pretty sure Eliezer thinks this heuristic should be applied to events that occurred in the past, not ones that will occur in the future—it’s a way of assessing whether a piece of evidence should be trusted or whether we should defy it. It’s also a way of weeding out hypotheses that don’t actually make experimental predictions. I don’t think he’s trying to say that we should ignore things that seem weird, particularly because he speaks out against the absurdity heuristic later on.
I’m pretty sure Eliezer thinks this heuristic should be applied to events that occurred in the past, not ones that will occur in the future—it’s a way of assessing whether a piece of evidence should be trusted or whether we should defy it. It’s also a way of weeding out hypotheses that don’t actually make experimental predictions. I don’t think he’s trying to say that we should ignore things that seem weird, particularly because he speaks out against the absurdity heuristic later on.