MarkovChain-Dot-Product-comparitor Alternatives

MarkovChain-Dot-Product-comparitor reviews and mentions

• OpenAI and Elon Musk
  2 projects | news.ycombinator.com | 6 Mar 2024

  Me.

  At least, I trust him as much as any other foreign[0] CEO.

  For all the stuff people complain about him doing, what I don't trust is that Californian ethics don't map well enough onto my ethics…

  …but I'm not sure any ethics works too well when examined. A while ago I came up with the idea of "dot product morality"[1] — when I was a kid, "good vs evil" was enough, then I realised there were shades of grey, then I realised someone could be totally moral on one measure (say honesty) and totally immoral on another (say you're a vegan for ethical reasons and they're a meat lover), and I figured you could say two people are "morally aligned" when their ethics vectors are pointing the same way.

  But more recently I realised that in a high dimensional space, there's a huge number of ways for vectors to be almost the same and yet 90° apart[2].

  [0] I'm not American, so to me he's foreign

  [1] I really need to move my blog to github, the only search results on Google are the previous times I posted this to HN: https://kitsunesoftware.wordpress.com/2019/05/25/dot-product...

  [2] Via trying to calculate the dot product of two Markov chains: https://github.com/BenWheatley/MarkovChain-Dot-Product-compa...