check in

algorithmic_intent.md

@@ -106,4 +106,4 @@ pretending to be stupid: https://slatestarcodex.com/2014/08/14/beware-isolated-d
 
 also address: machine learning, deception in nature—it's _really convenient_ to use the "learning" and "deception" codewords, and there's no convenient replacement—maybe that suggests that they are the right words
 
-https://www.greaterwrong.com/posts/DSnamjnW7Ad8vEEKd/trivers-on-self-deception/comment/CandwLBdJXXq7Qxet
+https://www.greaterwrong.com/posts/DSnamjnW7Ad8vEEKd/trivers-on-self-deception/comment/CandwLBdJXXq7Qxet

comment_on_articulating_lay_theories.md

@@ -0,0 +1,22 @@
+## Comment on "Articulating Lay Theories Through Graphical Models"
+
+In ["Articulating Lay Theories Through Graphical Models: A Study of Beliefs Surrounding Vaccination Decisions"](http://www.derekmpowell.com/publication/lay-theories-cogsci) (a conference paper from [CogSci 2018](https://cognitivesciencesociety.org/past-conferences/)), Derek Powell, Kara Weisman, and Ellen M. Markman
+
+
+
+
+7 point scale from "Strongly Agree" to "Strongly Disagree"
+
+"Natural things are always better than synthetic alternatives"
+"Parents should trust a doctor's advice even if it goes against their intuitions"
+
+r = .825, accounting for 68.1% of the variance in observed responses
+
+
+https://www.lesswrong.com/posts/hzuSDMx7pd2uxFc5w/causal-diagrams-and-causal-models
+
+trade-off between sparsity (more intelligible) and predictive
+
+[^crux]
+
+[^crux]: Thanks to [Anna Salamon](https://www.lesswrong.com/users/annasalamon) for this observation.