An accusation of privileging a hypothesis will be more persuasive if you also point out other families of hypotheses that together still deserve the majority of the probability mass.
But of course. It’s just my guess, given these data and personal experience, kinda like when Eliezer made a guess about procrastination. It’s the same guess that McCabe & Castel made.
The wording you used doesn’t reflect the extremely low probability. The hypothesis could be the best specific guess (which is still no good, just the best we have), and work as raw material for hypotheses that have more chance of actually capturing the situation (constructed by similarity to the first guess), but that can also be expressed by something like “my best guess is that something roughly like X might be happening”, instead of “I think X is happening”. If my best guess X is no good, I don’t think that X is happening.
Also, there probably should be a new standard fallacy on LW, “appeal to Eliezer”.
I updated my wording after your original comment on this topic. And I don’t agree that it’s probability is ‘extremely low’. I don’t think it’s the only explanation, merely that it’s often part of the explanation. It seems you’re taking me to be making a stronger claim than I’m intending to make.
My link to Eliezer’s post wasn’t meant to justify my practice, only to put it in context.
I don’t see how you’re justified in thinking that. It’s too detailed a hypothesis to locate using that data.
An accusation of privileging a hypothesis will be more persuasive if you also point out other families of hypotheses that together still deserve the majority of the probability mass.
But of course. It’s just my guess, given these data and personal experience, kinda like when Eliezer made a guess about procrastination. It’s the same guess that McCabe & Castel made.
The wording you used doesn’t reflect the extremely low probability. The hypothesis could be the best specific guess (which is still no good, just the best we have), and work as raw material for hypotheses that have more chance of actually capturing the situation (constructed by similarity to the first guess), but that can also be expressed by something like “my best guess is that something roughly like X might be happening”, instead of “I think X is happening”. If my best guess X is no good, I don’t think that X is happening.
Also, there probably should be a new standard fallacy on LW, “appeal to Eliezer”.
I updated my wording after your original comment on this topic. And I don’t agree that it’s probability is ‘extremely low’. I don’t think it’s the only explanation, merely that it’s often part of the explanation. It seems you’re taking me to be making a stronger claim than I’m intending to make.
My link to Eliezer’s post wasn’t meant to justify my practice, only to put it in context.