"Sexual Dimorphism": explaining high-dimensionality, again
[Ultimately_Untrue_Thought.git] / notes / sexual-dimorphism-in-the-sequences-notes.md
index 6afa647..933173c 100644 (file)
@@ -148,3 +148,30 @@ people like me being incentivized to identify as part of a political pressure gr
 
 Yudkowsky says trans rights!
 
+
+https://www.lesswrong.com/posts/jMTbQj9XB5ah2maup/similarity-clusters
+https://www.lesswrong.com/posts/cFzC996D7Jjds3vS9/arguing-by-definition
+
+https://www.lesswrong.com/posts/9QxnfMYccz9QRgZ5z/the-costly-coordination-mechanism-of-common-knowledge
+
+
+as demonstrated by the likes of [FaceApp](https://www.faceapp.com/), the _uniquely best piece of software in the world_.
+
+(Note that we can already basically do this for _images_ of female and male faces, using the [latent spaces found by generative adversarial networks]() and [autoencoders](), 
+
+Doing it for _actual whole people in the real world_ and not just flat images is a task for future superintelligences, not present-day GANs, but some of same basic principles should apply.)
+
+[TODO: Mathematically,  
+Joel et al. and response—maybe in next paragraph
+Beyond the Binary: https://www.pnas.org/content/112/50/15468
+http://cogprints.org/10046/1/Delgiudice_etal_critique_joel_2015.pdf
+
+
+> once you draw a boundary around a group, the mind starts trying to harvest similarities from the group. And unfortunately the human pattern-detectors seem to operate in such overdrive that we see patterns whether they're there or not; a weakly negative correlation can be mistaken for a strong positive one with a bit of selective memory.
+https://www.lesswrong.com/posts/veN86cBhoe7mBxXLk/categorizing-has-consequences
+
+96.8% classification from MRI https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6374327/ 
+]
+[the wrists: http://unremediatedgender.space/papers/yune_et_al-beyond_human_perception_sexual_dimorphism_in_hand_and_wrist_radiographs.pdf]
+
+Face editing with Generative Adversarial Networks: https://www.youtube.com/watch?v=dCKbRCUyop8
\ No newline at end of file