I plan to put all the acknowledgments and references in the last part; but I should mention now that I first read about the self-fulfilling prophecy problem on the decision-theory mailing list, in a post by Vladimir Slepnev. (EDIT: He passes on the credit to Benja Fallenstein in this LW comment.)
Other people have used “naive decision theory” before, some of them on Less Wrong, but none of the usages seemed to stick. So I called this one (which was an early candidate for UDT before the problem was noticed) Naive Decision Theory. I can change the name if people prefer.
I propose: Very Naive Decision Theory, as it represents an effort to make a decision theory which is maximally naive by making an assumption that the makers of naive decision theory are naive about the Godel’s incompleteness theorem and Halting problem, yet allow their decision theory to recurse rather than simply black-box it from itself. In practice, the black boxing is always done to cut down on the work because you are an applied mathematician, and you got a problem to solve, and you won’t enter the recursion if you can avoid doing that.
I first read about the self-fulfilling prophecy problem on the decision-theory mailing list, in a post by Vladimir Slepnev
As far as I know, the problem of self-fulfilling prophecies and spurious counterfactuals was first pointed out by Benja Fallenstein in the comments to one of my posts.
I plan to put all the acknowledgments and references in the last part; but I should mention now that I first read about the self-fulfilling prophecy problem on the decision-theory mailing list, in a post by Vladimir Slepnev. (EDIT: He passes on the credit to Benja Fallenstein in this LW comment.)
Other people have used “naive decision theory” before, some of them on Less Wrong, but none of the usages seemed to stick. So I called this one (which was an early candidate for UDT before the problem was noticed) Naive Decision Theory. I can change the name if people prefer.
Vote up this comment if I should stick with calling it Naive Decision Theory.
Vote up this comment if I should change the name of Naive Decision Theory. Reply with a suggestion or upvote one of the suggestions that’s been made.
I propose: Very Naive Decision Theory, as it represents an effort to make a decision theory which is maximally naive by making an assumption that the makers of naive decision theory are naive about the Godel’s incompleteness theorem and Halting problem, yet allow their decision theory to recurse rather than simply black-box it from itself. In practice, the black boxing is always done to cut down on the work because you are an applied mathematician, and you got a problem to solve, and you won’t enter the recursion if you can avoid doing that.
As far as I know, the problem of self-fulfilling prophecies and spurious counterfactuals was first pointed out by Benja Fallenstein in the comments to one of my posts.
KARMA BALANCE: Call it “This Must Be The Mathematical Universe (Naive Decision Theory)”!