As I mentioned in a different thread, I want to study both Maths and Economics because I don’t know whether I will want to go into stats stuff or policy stuff, but I now think that this is not as important as I thought it was.
Transient Altruist
Karma: 101
Yeah my worry was that I don’t know whether I wanted to go into stats stuff or policy, so I thought that if I just did Maths at Oxford I wouldn’t gain enough information about which is a better fit for me to be able to make the switch into economics, which is possible but requires a commitment.
Having reconsidered, I think there are likely to be ways of learning which route I want to go down even if I don’t go to a uni where I can study both, and the costs of doing this are likely to be outweighed by the benefit of attending Oxford, especially considering the responses this post has gotten.
(This means I am now unlikely to apply to American universities since I am able to attend a top uni in the UK, and the probability of getting accepted to an American one AND not Oxford is even lower [especially since the American assessment criteria suit me worse])
Sorry there’s something really basic I don’t get about Example 1. There is a good chance I am mistaken but I think this confusion will be shared by many so I think it’s worth commenting.
Your point is that Scenario 1 is what they would come to if they each try to maximise their personal counterfactual impact. In your example, Alice and Bob are calculating their counterfactual impact in each scenario on the basis that the other person’s decision is not affected by theirs. But if both are trying to maximise their personal counterfactual impact, then the other person’s decision IS affected by theirs! So their calculations of counterfactual impact were wrong!
Each person can say to the other “I’ll donate to P if you donate to P, otherwise I’ll donate to Q/R” (because that is what would maximise their counterfactual impact). Each will then see that donating to P provides a counterfactual impact of 5 (utility of 15 is created rather than 10), while donating to Q/R gives a counterfactual impact of 10 (20 is created rather than 10). They will both donate to Q/R, not P like you suggest.
(You CAN make it so that they can’t communicate and they each THINK that their decision doesn’t affect the other’s [even though it does], and this WOULD make the counterfactual impact in each scenario the same as the ones you give. BUT you would have to multiply those impacts by the probability that the other person would make the given choice, and then combine it with the situation where they don’t… so it doesn’t work the same and even then it would be an information issue rather than a problem with maximising counterfactual impact...)
Well I’m currently taking Economics A-level and I do find it really interesting so I can see myself pursuing economics in the future, but I don’t plan on it specifically. That’s why I thought it might be good to be able to do a bit of it to know whether to do it at masters? Maybe the A-level is enough though
Thanks!
Oh I had assumed that was more of a secondary consideration because of being hard to predict well. I’ll make sure to talk to some more people.
The only modules you can’t do at Oxford are the Economics ones. I hear it’s relatively easy to switch into Economics from Maths, but maybe it would be good to be able to try it out earlier?
I also got the same feeling, but then discarded it because this is not supposed to be a prioritisation argument, simply motivation.
It doesn’t need to (suspiciously) claim that AI safety so happens to also be best for your other interests, just that it helps there too, and that that’s nice to know :)
So long as make your commitments based on solid rational reasoning, it’s ok to lean into sources of motivation that wouldn’t be intellectually persuasive, but motivate you nonetheless.
Ah yes I forgot about that. Considering I’m personally equally able/willing to do Maths, things seem to be lining up to say that Maths gives much of the same benefits as Econ but keeps more options open.
The thing is I really don’t have any specific objectives/goals. It sounds like Maths is best for keeping options open, so I suppose I should do that and narrow down later.
With that being the case, do you think there are even any benefits to doing Economics from the start rather than Maths?
Great point! Especially considering I’m already a believer in the multiverse, it’s not necessarily less likely that we’ll develop some method of creating infinite utilities than it is that heaven or hell exists.
I’ll definitely consider this, so thanks for commenting!
Yes this is exactly what I’m saying
The only point I was making was that not all versions of God are equally likely, so the possible utilities of heaven and hell don’t cancel. I don’t know what the most likely form of God is, but it sounds like we both agree that not all of them are equally likely.
I clarified in my edit at the top of my post what I mean by “accept Pascal’s Wager”. To repeat I see it as accepting the idea that way to do the most (expected) good is to prevent as many people as possible from going to hell, and cause as many as possible to go to heaven, regardless of how likely it is that heaven/hell exists (as long as it’s non-zero).
As for what this entails I have no idea. For now I’m just trying to decide whether to pursue this aim or not. The way I would actually do that comes later, if I choose to accept.
Oh wait sorry I got confused with totally different comment that did add an extra assumption. My bad...
As for the actual comment this thread is about, expected value theory can be derived from the axioms of VNM-rationality (which I know nothing about btw), whereas proposition 3 is not really based on anything as far as I’m aware, it’s just a kind of vague axiom of itself. I feel we should restrain from using intuitions as much as possible except when forced to at the most fundamental level of logic — like how we don’t just assume 1+1=2, we reduce it to a more fundamental level of assumptions: the ZFC axioms.
In summary, propositions 1 and 3 are mutually exclusive, and I think 1 should be accepted more readily due to it being founded in a more fundamental level of assumptions.
I suppose I could see reason to make this assumption, given that you could get used to the luxuries of heaven and it would start to be less pleasurable. However this doesn’t really eliminate the problem because there’s still the possibility that this assumption is incorrect, meaning the probability of infinite payoff is still non-zero and therefore the wager still stands.
That’s true but I think we need to make the least number of intuition based assumptions possible. Yitz’s suggestion adds an extra assumption ON TOP of expected value theory, so I would need a reason to add that assumption.Oops I got mixed up and that response related to a totally different comment. See my reply below for my actual response
This argument is one that makes intuitive sense, and of course I am no exception to that intuition. However intuition is not the path to truth, logic is. Unless you can provide a logic-founded reason why almost certain loss with a minuscule chance of a huge win is worse than unlikely loss with a probable win, then I can’t accept the argument.
Thanks, I appreciate the explanation.
I think the tone I wrote it in gave the impression that I was saying “I don’t see how what they did is illegal” rather than “what did they do that was illegal”. I try and avoid that kind of thing next time.
I asked this question on the EA Forum just because EA is why I’m interested in the story, and I trust the EA Forum to produce good discussion that gets to heart of what I’m interested in, compared to the only other place where I know I can ask a direct question, which is Reddit. Any implication that there was still a mystery to be figured out was accidental.
As for the half-baked takes—actually I don’t think I gave any. None of what I said was an opinion it was simply an account of the easily findable facts.