Good heavens. Have philosophers never read about how Bayes nets work? Uncertain observations are technology at this point; you send up a non-extreme lambda message. (And this doesn't even violate probability theorems qua theorems, as so many approximations understandably do.)
-
-
And I say this to illustrate a larger dichotomy: naive toolboxers see a problem and think they've discovered a context in which to not use that tool. Sophisticated thinkers who have Lawful thinking as an option are much more likely to wonder if the generalization still holds.
1 reply 0 retweets 0 likes -
Replying to @ESYudkowsky @ArtirKel
“Naive Toolboxer” seems analogous to Straw Vulcan here.
1 reply 0 retweets 1 like -
Replying to @Meaningness @ArtirKel
Absolutely. Sophisticated thinkers can conceive of both context-dependent recipes and universal generalizations. They are ready to adapt tools as required, and expect to repair laws without compromising them. But Straw Vulcans exist in real life, and so do naive toolboxers.
1 reply 0 retweets 2 likes -
Replying to @ESYudkowsky @ArtirKel
Okay… we agree that it is possible to fail to apply rational methods when they would be useful, and some people may make that mistake often.
1 reply 0 retweets 1 like -
Replying to @Meaningness @ArtirKel
That's still Toolbox thinking! There are perspectives on life besides whether some recipe would be useful at a given moment!
1 reply 0 retweets 4 likes -
Replying to @ESYudkowsky @ArtirKel
Yes, you are right. Let me amend the previous tweet to “framework” from “method.” This may be revealing, as you are arguing. 1/2
1 reply 0 retweets 1 like -
I do understand the object-level distinction between mathematical frameworks and specific methods. However, at the meta-rational level, frameworks are themselves just methods, because they are not universally applicable. 2/3
1 reply 0 retweets 1 like -
So the object-level distinction between frameworks and methods is not as salient for me as it is for you. 3/3
1 reply 0 retweets 1 like -
Replying to @Meaningness @ArtirKel
One factor distinguishing the Lawful metatool is the expectation of hidden, nonobvious simplicity and generality. If Jaynes had possessed *more* faith in probability, he might have been more likely to extend it to non-omnisicient distributions on quantified sentences.
2 replies 0 retweets 2 likes
Well, that’s very interesting, and makes sense! I definitely do not have an expectation of hidden simplicity. Perhaps that fundamental prior is what distinguishes our worldviews.
-
-
Replying to @Meaningness @ArtirKel
My own big contribution to decision theory was logical decision theory, which replaced a huge number of complicated patches to the "fatal" objection of Newcomb's Problem with a simple, unified view. Because I knew it couldn't possibly *actually* be the case that the...
1 reply 0 retweets 2 likes -
...basic decision principle was this big weird loopy structure containing lots of exceptions and complications and where the meta-level wasn't consistent with the object level and so on. I knew that had to be humans making mistakes and missing the obvious.https://arbital.com/p/logical_dt/
1 reply 0 retweets 4 likes - 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.