January 1, 2021
Things I’ve Changed My Mind About
- I used to dismiss signaling as a way to explain behavior, now I think it has significant explanatory power for many aspects of human behavior.
- I used to think flossing was pointless, now I do it everyday. I think this was case of confusing “there’s no evidence this works” with “there’s evidence that this doesn’t work”.
- I thought Twitter doubling it’s character limit would be terrible, but now I think it was actually great.
- I didn’t used to think comparing AGI to corporations was a particularly useful analogy, now I think the comparison between ‘ruthlessly optimizing for an objective’ and profit-maximizing is very apt and that we can learn from the mechanisms we use to control corporations when it comes to building safe AGI.
- I used to be very anti-drugs, now I am mostly in favor of legalizing most drugs.
- I was brought up Catholic and I am now an atheist.
- I used to think there wasn’t much point in reading the classics (because I had pretty much absorbed the storylines/messages from other aspects of culture, and because it seemed like a waste of time to read the same thing as everyone else), then I read some classics and now I realize there’s a reason they are classics.
- I used to believe very strongly in the power of the placebo effect but now I am less certain or at least, I think the placebo effect may be much weaker than we thought, due to things like the Hawthorne effect and Simpson’s paradox. More here.
- I feel much more favorably towards qualitative research methods than I used to.
- I feel much more favorable towards markets as a way of allocating resources than I used to.
- Covid-19 has made me more aware of the risks of some things I am generally very in favor of, such as:
- High-density living
- Open borders / increased travel / immigration
- Relying on global supply chains
- I used to be a big believer in Implicit Association Testing, I’m now much less confident it actually measures anything meaningful. Similarly, I used to be proponent of unconscious bias training, but it seems like that might actually be counterproductive.