X-Git-Url: http://unremediatedgender.space/source?p=Ultimately_Untrue_Thought.git;a=blobdiff_plain;f=notes%2Fnotes.txt;h=c81c9856e66403e49f31115d60ba58541583b7b1;hp=203e86cf77b90015050965ec9288b5adc3829370;hb=3bec30b88547dc92385bcd7534a6e4cc1825cfd2;hpb=f6344c10707581308aff24e62f0878f98a5c90fd

diff --git a/notes/notes.txt b/notes/notes.txt
index 203e86c..c81c985 100644
--- a/notes/notes.txt
+++ b/notes/notes.txt
@@ -2934,3 +2934,17 @@ https://www.glowfic.com/posts/4508?page=21
 https://www.reddit.com/r/TheMotte/comments/ruvu1k/culture_war_roundup_for_the_week_of_january_03/hrdqmzk/
 
 https://letter.wiki/conversation/1232
+
+----
+
+in the context of quantifying predictions, [in the post evaluating his 2020 predictions, Alexander writes](https://astralcodexten.substack.com/p/2020-predictions-calibration-results):
+
+> We have a debate every year over whether 50% predictions are meaningful in this paradigm; feel free to continue it.
+
+Someone reading this who trusted Alexander as a general-purpose intellectual authority ("the best of us", the "rationalists") might walk away with the idea that it's an open problem whether 50% binary predictions are meaningful—perhaps reasoning, if the immortal Scott Alexander doesn't know, then who am I to know?
+
+But it's not. On this website, [Rafael Harth explains why 50% isn't special](https://www.lesswrong.com/posts/DAc4iuy4D3EiNBt9B/how-to-evaluate-50-predictions). I claim that this should actually be pretty obvious to competent quantitative thinkers, even if it's not obvious to the collective _SSC_/_ACX_ commentariat, and Alexander can't tell which of his commenters are competent quantitative thinkers.
+
+I don't particularly fault Scott for this: [by his own admission, he's not a math guy](https://slatestarcodex.com/2015/01/31/the-parable-of-the-talents/). (And the vast majority of math people can't write as well or as fast as Scott. No one is the best at everything!) Rather, I'm saying that a culture that wants to _actually_ be right about everything would do better to _just_ focus on being right on the object level, without [wireheading on its own promises of being right about everything](http://benjaminrosshoffman.com/effective-altruism-is-self-recommending/).
+
+(Incidentally, Scott himself is actually very good about [not trying to claim more authority than is actually justified by his performance](https://slatestarcodex.com/2019/07/04/some-clarifications-on-rationalist-blogging/). His fans should try to be more like him along this dimension!)