Friday, April 24, 2015

Topics I need to discuss

I have discussing quite a few topics on my todo list, and I probably should periodically create short lists of what I haven't discussed. This is one of them.
  • Immortality, cloning, and having kids. I don't want to be myself forever...
  • The problem of language. "Monkey see. Monkey do." is a completely different paradigm from "Simon says," and "Simon says" is worse.
  • Cryonics. I'm in favor, but mostly as signalling.
  • Death. If it's not terminal bad, it's instrumental good.
  • Error theory. Sort of. God is the tribe personified, and morality is the will of god.
  • More error theory. I'm much more sympathetic to the idea of the existence of meta-level claims about morality that are true, than I am about the possibility of the existence of true object-level claims about morality.
  • Utilitarianism, now and later. How does the fact that what people value changes over time fit with utilitarianism?
  • Utilitarianism: present and future. Trying to optimize future utility is very different from trying to optimize present utility.
  • The null hypothesis -- How do we decide which hypothesis is null and is there a better way to do it?
  • Privileging the hypothesis. I don't think it's as bad as Eliezer does.
  • An alternative to Occam's Razor, genetic algorithms based on modifying existing hypotheses get to start with probability stolen from their parents.
  • Everyone is a sociopath with respect to their enemies. Not being a sociopath is just about belonging to some in-group.
  • Sometimes I wonder if I'm on the autism spectrum. My approach to morality, the way I review books, and the particular ways I like to approach organization all seem a bit autistic to me. In a good way, I think.
  • To be verbs -- are they bad?
  • Solomonoff Induction -- is it actually asymptotically optimal?
  • Should I seek to be happy?
    • Talk about sanity...
  • Humor, hyperbole, and sarcasm. Should I try to be funny?
  • Slate Star Codex: No arbitrage in social science but plenty in business, even though businesses are subject to much higher selective pressure. Hmm...
  • Slate Star Codex on the control group being out of control, and related. How should you update your beliefs as you collect new evidence?
  • Time, evidence, and over-fitting. Only predictions provide evidence for the accuracy of a claim. Well, only predictions and explanatory power.
  • Time, evidence, and equations. How would you expect physics to be different if we are in a timeless universe versus if we are in a universe where time is relevant?
  • Models of science need to include models of scientists.
  • An introduction to extreme skepticism.
  • Be wary of non-monotonic trends.
  • Self-perpetuating phenomena and multi-faceted explanations of major shifts.
  • What are reasonable "prior" odds?
  • Andy Warhol sponsoring art.
  • A quick overview of social history as I understand it.
  • Evolution: an introduction to the code I'm writing to test ideas followed up with occasional posts about results.
  • Pronouns. Easy reference please. (Let's have more than one pronoun even if we have gender neutral pronouns.)
  • Language is an encryption.
  • City states have small governments. Empires have big governments.
  • I was wrong about Lawful Good and literature. Terry Pratchett totally does it.
  • Resource allocation is hard!

No comments:

Post a Comment