Failure mode: My “something to protect” is to spread rationality throughout the world and to raise the sanity waterline, which is best achieved by having my own rationality dojo.
I agree. I think that failure mode might then be better avoided by restricting possible “somethings”, as opposed to adding another requirement on to one’s reasons for wanting to be rational.
Yes, but that’s an exercise implicitly left to the reader. Formulating it this way is somewhat intuitively easier to understand, and if you’ve read the other sequences this should be simple enough to reduce to something that pretty much fits (restriction of “things to protect”) in beliefspace.
Essentially, this article, the way I understand it, mostly points at an “empirical cluster in conceptspace” of possible failure modes, and proposes possible solutions to some of them, so that the reader can deduce and infer the empirical cluster of solutions to those failure modes.
The general rule could be put as “Make rationality your best means, but never let it become an end in any way.”—though I suspect that I’m making a generalization that’s a bit too simplistic here. I’ve been reading the sequences in jumbled order, and I’m particularly bad at reduction, which is one of the Sequences I haven’t finished reading yet.
Failure mode: My “something to protect” is to spread rationality throughout the world and to raise the sanity waterline, which is best achieved by having my own rationality dojo.
Beware the meta.
I agree. I think that failure mode might then be better avoided by restricting possible “somethings”, as opposed to adding another requirement on to one’s reasons for wanting to be rational.
Yes, but that’s an exercise implicitly left to the reader. Formulating it this way is somewhat intuitively easier to understand, and if you’ve read the other sequences this should be simple enough to reduce to something that pretty much fits (restriction of “things to protect”) in beliefspace.
Essentially, this article, the way I understand it, mostly points at an “empirical cluster in conceptspace” of possible failure modes, and proposes possible solutions to some of them, so that the reader can deduce and infer the empirical cluster of solutions to those failure modes.
The general rule could be put as “Make rationality your best means, but never let it become an end in any way.”—though I suspect that I’m making a generalization that’s a bit too simplistic here. I’ve been reading the sequences in jumbled order, and I’m particularly bad at reduction, which is one of the Sequences I haven’t finished reading yet.