In many cases, it seems very doubtful whether further research into whether animals are conscious will be action-guiding in any meaningful way. Further research into whether chickens are conscious, say, will not produce definitive certainty that they are or aren't. (Finding out whether your favorite theory of consciousness predicts that they are conscious is only so useful, as we should have significant uncertainty about the right theory of consciousness.) And moderate changes in your credence probably shouldn't affect what you should do. E.g., if your credence in chicken consciousness drops 20%, there is still the moral risk argument for acting as if chickens are conscious; if you have some reason for rejecting that argument, that was probably also a reason when your credence was 20% higher. And at the same time, there are potentially very great opportunity costs to waiting to act--costs that don't make sense if decision-relevant information isn't actually likely to come in.

The EarlyModernTexts versions are what I read when I had to do my comprehensive history exam in grad school. I recommend them.

I accept the bullet biting response. I think someone who doesn't should say the utility of the observers may outweigh Jones' utility but that you should save Jones for some deontic reason (which is what Scanlon says), or maybe that many small bits of utility spread across people don't sum in a straightforward way, and so can't add up to outweigh Jones' suffering (I think this is incorrect, but that something like it is probably what's actually driving the intuition). I think the infinite disutility response is wrong, but that someone who accepts it should probably adopt some view in infinite ethics according to which two people suffering infinite disutility is worse than one--adopting some such view may be needed to avoid other problems anyway.

The solution you propose is interesting, but I don't think I find it plausible:

1. If Jones' disutility is finite, presumably there is some sufficiently large number of spectators, X, such that their aggregate utility would outweigh his disutility. Why think that, in fact, the physically possible number of observers is lower than X?

2. Suppose Jones isn't suffering the worst torment possible, but merely "extremely painful" shocks, as in Scanlon's example. So the number of observers needed to outweigh his suffering is not X, but the lower number Y. I suppose the intuitive answer is still that you should save him. But why think the physically possible number of observers is below Y?

3. Even if, in fact, the physically possible number of observers is lower than X, presumably the fundamental moral rules should work across possible worlds. And anyway, that seems to be baked into the thought experiment, as there is in fact no Galactic Cup. But even if the physically possible number of observers is in fact lower than X, it could be higher than X in another possible world.

4. Even if the possible number of observers is in fact finite, presumably there are possible worlds with an infinite number of possible observers (the laws of physics are very different, or time is infinite into the future, or there are disembodied ghosts watching, etc.). If we think the solution should work across possible worlds, the fact that there can only be a finite number of observers in our world is then irrelevant.

5. You assume our lightcone is finite "with certainty." I assume this is because of the expected utility concern if there is some chance that it turns out not to be finite. But I think you shouldn't have epistemic certainty that there can only be a finite number of observers. 

6. The solution seems to get the intuitive answer for a counterintuitive reason. People find letting Jones get shocked in the transmitter case counterintuitive because they think there is something off about weighing one really bad harm against all these really small benefits, not because of anything having to do with whether there can only be a finite number of observers, and especially not because of anything having that could depend on the specific number of possible observers. Once we grant that the reason for the intuition is off, I'm not sure why we should trust the intuition itself.

*I think your answer to 1-3 may be that there is no set-in-stone number of observers needed to outweigh Jones' suffering: we just pick some arbitrarily large amount and assign it to Jones, such that it's higher than the total utility possessed by however many observers there might happen to be. I am a realist about utility in such a way that we can't do this. But anyway, here is a potential argument against this:

Forget about what number we arbitrarily assign to represent Jones' suffering. Two people each suffering very slightly less than Jones is worse than Jones' suffering. Four people each suffering very slightly worse than them is worse than their suffering. Etc. If we keep going, we will reach some number of people undergoing some trivial amount of suffering which, intuitively, can be outweighed by enough people watching the Galactic Cup--call that number of observers Z. The suffering of those trivially suffering people is worse than the suffering of Jones, by transitivity. So the enjoyment of Z observers outweighs the suffering of Jones, by transitivity. And there is no reason to think the actual number of possible observers is smaller than Z.

Different ways of calculating impact make sense in different contexts. What I want to say is that the way Singer, MacAskill, GiveWell are doing it (i) is the one you should be using in deciding whether/where to donate (at the very least assuming you aren't in some special collective action problem, etc.) and (ii) one that is totally fine by ordinary standards of speech--it isn't deceptive, misleading, excessively imprecise, etc. Maybe we agree.

The criticism of the concept of "effective altruism," and the second main criticism to the extent that it's related to it, also feels odd to me. Altruism in the sense of only producing good is not realistically possible. By writing your Wired article, say, it is overwhelmingly likely that you set off a chain of events that will cause a different sperm to fertilize a different egg, which will set off all sorts of other chains of events meaning that in hundreds of years we will have different people, different weather patterns, different all sorts of stuff. So writing this will cause untold deaths, untold suffering, that wouldn't have occurred otherwise. So too for your friend Aaron in the Wired article helping the people on the island, and so too for anything else you or anyone might do.

So either altruism is something other than doing only good, or altruism is impossible and the most we can hope for is some kind of approximation. It wouldn't follow that maximizing EV is the best way to be (or approximate being) altruistic, but the mere fact that the actions EAs take are like all other actions in that they have some negative consequences is not in itself much of a criticism.

The first criticism feels pretty odd to me. Clearly what Singer, MacAskill, GiveWell, etc. are talking about is the counterfactual impact of your donation, since that is the thing that should be guiding your decision-making. And that seems totally fine and in accordance with ordinary English: it is fine to say that I saved the life of the choking person by performing the Heimlich maneuver, that Henry Heimlich saved X number of lives by inventing the Heimlich maneuver, that my instructor saved Y number of lives by teaching people the Heimlich maneuver, and that a government program to promote knowledge of the Heimlich maneuver saved Z lives, even if X + Y + Z + 1 is greater than the overall number of lives saved by the Heimlich maneuver. And if I were, say, arguing for increased funding of the government program by saying it would save a certain number of lives, it would be completely beside the point to start arguing that I should actually divide the marginal impact of increased funding to account for the contribution of Henry Heimlich, etc.

They are separate views, but related: people with person-affecting views usually endorse the asymmetry, people without person-affecting views usually don't endorse the asymmetry, and person-affecting views are often taken to (somehow or other) provide a kind of justification for the asymmetry. The upshot here is that it wouldn't be enough for people at OP to endorse person-affecting views: they'd have to endorse a version of a person-affecting view that is rejected even by most people with person-affecting views, and that independently seems gonzo--one according to which, say, I have no reason at all not to push a button that creates a trillion people who are gratuitously tortured in hell forever.

Very roughly, how this works: person-affecting views say that a situation can't be better or worse than another unless it benefits or harms someone. (Note that the usual assumption here is that, to be harmed or benefited, the individual doesn't have to exist now, but they have to exist at some point.) This is completely compatible with thinking it's worse to create the trillion people who suffer forever: it might be that their existing is worse for them than not existing, or harms them in some non-comparative way. So it can be worse to create them, since it's worse for them. And that should also be enough to get the view that, e.g., you shouldn't create animals with awful lives on factory farms.

Of course, usually people with person-affecting views want it to be neutral to create happy people, and then there is a problem about how to maintain that while accepting the above view about not creating people in hell. So somehow or other they'll need to justify the asymmetry. One way to try this might be via the kind of asymmetrical complaint-based model I mentioned above: if you create the people in hell, there are actual individuals you harm (the people in hell), but if you don't create people in heaven, there is no actual individual you fail to benefit (since the potential beneficiaries never exist). In this way, you might try to fit the views together. Then you would have the view that it's neutral to ensure the awesome existence of future people who populate the cosmos, but still important to avoid creating animals with net-negative lives, or future people who get tortured by AM or whatever.

Now, it is true that people with person-affecting views could instead say that there is nothing good or bad about creating individuals either way--maybe because they think there's just no way to compare existence and non-existence, and they think this means there's no way to say that causing someone to exist benefits or harms them. But this is a fringe view, because, e.g., it leads to gonzo conclusions like thinking there's no reason not to push the hell button.

I think all this is basically in line with how these views are understood in the academic literature, cf., e.g., here.