https://www.effectivealtruism.org/articles/cause-profile-lon...
Or to put it another way, everything fuzzes out into noise for me much sooner than humanity will have trillions of new members. There's no way for me to predict whatsoever what effect any action I take today will have a thousand years from now. Even in extreme cases, like, I push a magic button that instantly copies whatever you, the reader, believe is the optimal distribution of ideological beliefs out into the world (ignoring for the moment the possibility that your ideology might consider that unethical, this is just a thought experiment anyhow so no need to go that meta), you really don't know what that would do 1000 years from now, what the seventeenth-order effects of such a thing would be. I'm not even saying that it might not be as good as you think or something; I'm saying you just have no idea what it would be at all. So there's no way to hold people responsible for that, and no way to build plans based on it.
I think the core idea is simply: since resources for helping the poor/sick is not unlimited, we should try to allocate those resources in the most effective way. Before EA charity evaluation came along, the only metric for most people was simply looking at the charity overhead via Charity Navigator. But that isn't a great metric. A charity with only a 1% overhead with a mission to make balloon animals for children dying in a famine will score well on Charity Navigator but does nothing to help the problem.
To be honest I haven't looked deeply into long-termism, but from what I've heard (eg, hearing Will MacAskill on a few podcasts) it seems to ignore a few things. Just like a bird in the hand is worth two in the bush, long-termers have no good way to estimate the likelihood of future events, and that discounting needs to increase greatly the further out one looks. At best many of these estimates are like the Drake Equation -- better than nothing, but with multiple orders of magnitude error bars.
There are other second-order reasons which don't seem to factor in, or at least haven't come across in the few hours of listening to long-termers talk about the issue. One is that by working to make a better world now, it effects the trajectory of future events much more directly than the low-probability guesswork they think my have an impact in the distant future.
"The great subverter of Pyrrhonism [radical skepticism] is action, and employment, and the occupations of common life. [...] I dine, I play a game of back-gammon, I converse, and am merry with my friends; and when after three or four hour's amusement, I wou'd return to these speculations, they appear so cold, and strain'd, and ridiculous, that I cannot find in my heart to enter into them any farther."
Hume
The biggest positive change you can make, even for future generations, is to uplevel the people who are alive today.
Although we are at a peak population of a bit over 8B people at the moment, it is estimated that more than 100B people we would classify as humans have ever lived. The population long ago was much smaller than 1B, but thousands of generations have lived and died.
Likewise, the guy who invented the Segway probably thought he was going to revolutionize personal transportation, but that didn't pan out (yet, at least).
2. Cover up the chart so only the data from the past to present day is visible.
3. Note that most humans in that subset exist near or at the present. You are one of these people today, it should make sense for you to be born in one of the densest parts of the graph.
4. Now uncover the graph. If there are trillions of humans in the future, it seems almost impossibly unlikely that you would born in a part of the graph with "so few" humans as today, and not in the far future.
Therefore, you must conclude that the actual graph rapidly drops to zero in the near future. QED.
This "doomsday argument" is a pretty shit one, but not worse than others I've seen arguing the opposite.