- If your only measure of progress is to compare the past to the present, then it’s easy to imagine you’re making progress when in reality you’re on a random walk
- One counterargument to the random-walk argument against morality is to say that the future will be more moral than the present, just as the present was more moral than the past
- But then aren’t we just extrapolating?
- Can you actually imagine a being that is more moral than yourself – one that believes that something you consider right and just is actually morally wrong?
- It’s easy to come up with examples of moral progress, but it’s much more difficult to show directionality and explain how directionality is implemented
- People are the product of evolution
- So why are people so nice?
- People do all sorts of things that would seem to be evolutionary disadvantageous
- How do we explain this without postulating some kind of shadowing figure dictating evolution?
- Our aesthetic and moral senses are a result of evolution, the same as our capacity for cruelty and destruction
- It is not a physical miracle that humans turned out to be as moral as they are, but a moral miracle
- It is entirely possible to recognize the kindness of humanity while recognizing the evolutionary origins of that kindness
- Just like life, at one point, arose from non-living matter, so too did creatures with a sense of morality arise from creatures that did not have a sense of morality
- Eliezer, in 1999, thought that building a generic superintelligence would be sufficient to “solve” morality
- Unfortunately, there is no reason to think that a superintelligence would necessarily have to think about morality
- You can’t compute morality without having some kind of starting point of what is and is not moral
- You can’t completely discard all products of evolution when thinking about morality, because that would involve discarding your own brain
- We should be willing to realize that we know at least a little bit about morality, and that little bit can form a starting point from which we improve
- We have to drop the notion that morality is something that “ghost of perfect emptiness” would agree with – to such an entity, even the question of what is and is not moral is meaningless
- Most existential angst isn’t existential
- It’s easy to think that life is inherently unhappy when you’ve given up on solving the problems that are making you unhappy
- The opposite of happiness isn’t sadness, it’s boredom
- Most feelings of existential angst can be blamed on people having at least one problem in their lives that they’ve given up on solving
- It’s an interesting question to wonder whether counterfactuals can be true
- We never actually experience counterfactuals
- Nor can we go back in time and replay events with different starting conditions
- So, given that, can we say anything in a counterfactual? Even things that are beyond physical causation?
- If we have a “lawful” computational procedure, the outcome of the counterfactual will be the result of that procedure on different starting conditions
- In order to compute counterfactuals, take some of the variables in your causal model, set them to different values, and then compute the final probability distribution
- Any philosophical work that takes counterfactual distributions as given without telling you how to compute those counterfactuals is largely useless