+
+The problem with uncritically validating an autodidactic's ego, is that a _successful_ autodidact needs to have an accurate model of how their studying process is working, and that's a lot harder when people are "benevolently" trying to _wirehead_ you.
+
+The man is so egregiously sexist in any other context, _except_ when I need the right answer to make extremely important medical decisions
+
+I don't need to be a mind-reader about how it feels because I can read, I can point to the text from 2010 and 2016+, and notice the differences
+
+Maybe he was following the same sucking-off-Stalin algorithm internally the whole time (just like Anna was), and I only noticed it in 2016+ because the environment changed, not the algorithm? I doubt it, though (rather, the environment changed the algorithm)
+
+http://benjaminrosshoffman.com/honesty-and-perjury/
+
+
+origins of the strawberry example
+https://www.facebook.com/yudkowsky/posts/10155625884574228
+https://www.facebook.com/yudkowsky/posts/10154690145854228
+(Note: we can do this with images of strawberries! The hard problem is presumably inventing nanotech from scratch.)
+
+lack of trust as a reason nothing works: https://danluu.com/nothing-works/ shouldn't the rats trust each other?
+