Recently I learned I wasn’t a longtermist because I don’t really buy our ability to forecast more than about 5 years in the future. So sure, I’d like to help future people but I’ve got no idea how, short of us not killing ourselves in the next 5 years. Supposedly this makes me not actually a longtermist???
You probably agree with me that (a) we can’t know whether it will rain on 2/10/2050 and (b) we can be pretty sure that there will be a solar eclipse on 7/22/2028. You are actively participating in a prediction market, so you seem to believe in some ability to forecast the future better than a magic 8 ball.
Where do you think the limits are to what kinds of things we can make useful predictions about, and how confident those predictions can be?
Yeah I guess the limits come down to the actions of others. I can’t see any human effecting the eclipse but I can see humans negating or misdirecting any actions I take toward the future. And the more humans there are, the more likely that becomes. Like, the more people there are the more likely any little trickle of a river you try to send into the future via society.. gets stepped on blocked or muddied somehow
I think that makes you a longtermist though… having read What We Owe the Future anyway, unless I missed something:
I think a longtermist would say that the effects on future moral people should dominate our moral calculus due to their vast number, not necessarily that they can right now. But we should keep an eye out for how to impact the longrun future positively and take such a chance if we ever see it. Some people think they see the chance now, so they are taking it.* Maybe you will never see something plausible-to-you within your lifetime. But that doesn’t mean a chance will never occur.
For example if we could run amazing simulations to test longrun outcomes, I think a longtermist, if they believed in the tech, would want to give the best-predicted longterm action a go (like, the best sum of experiences had over time, summed at the end), using neutral moral weights for beings living today vs 3 years from now vs 1000 years from now. Of course there will be wider ranges and larger confidence intervals but you’d factor those too, as to what option you’d want to do. By contrast a neartermist would add moral weight to the consequences and experiences for beings existing in the nearterm on top of the differing ranges and confidence intervals which are just sensible to use for both neartermists and longtermists.
*I’ll note that some extinction risks seem low enough percentage chance of happening to me that they may not be worth working on unless you do add neutrally-weighted future generations into the calculus. But it depends on the moral discount rate you’d use, like if your expected rate of population growth is large if everything goes well, and your moral discount rate is gradual enough, you might still end up preferring a “longtermist” intervention, even though you might not really be a true “longtermist” philosophically because you are still claiming that future generations are morally worth less (regardless of confidence), but focusing on the longterm just passed your bar anyway because of the scale.
Longtermism includes the claim that improving the future is tractable. I think that was probably a mistake, and it should just be a claim about values.
Daaaang yeah that seems wrong to me too unfortunately. I can imagine we have passed the threshold on coordination and technology that changing the long-run future somewhat predictably is a tractable cause now (and I can also imagine we haven’t), but I’m pretty sure there were many years in history where it would have been impossible to predict past 5 years ahead. It seems to me that a philosophical position (which longtermism claims to be) should be able to have existed then too. But if it included tractability, that position was likely impossible to hold (correctly) at some moments in history, or in many single-actor thought experiments. But I’m no philosopher
I am unsure if that makes you not a longtermist. My understanding of the basic longtermist claim is as follows:
Future individuals matter as much as those in the present
We should do something to positively shape their lives
I think this has little to do with forecasting accuracy, albeit if we were really good at that, it would help us better outline the something in the second point. I think the major part of it is about how we assign moral value to those who will exist in the future.
Recently I learned I wasn’t a longtermist because I don’t really buy our ability to forecast more than about 5 years in the future. So sure, I’d like to help future people but I’ve got no idea how, short of us not killing ourselves in the next 5 years. Supposedly this makes me not actually a longtermist???
You probably agree with me that (a) we can’t know whether it will rain on 2/10/2050 and (b) we can be pretty sure that there will be a solar eclipse on 7/22/2028. You are actively participating in a prediction market, so you seem to believe in some ability to forecast the future better than a magic 8 ball.
Where do you think the limits are to what kinds of things we can make useful predictions about, and how confident those predictions can be?
Yeah I guess the limits come down to the actions of others. I can’t see any human effecting the eclipse but I can see humans negating or misdirecting any actions I take toward the future. And the more humans there are, the more likely that becomes. Like, the more people there are the more likely any little trickle of a river you try to send into the future via society.. gets stepped on blocked or muddied somehow
I think that makes you a longtermist though… having read What We Owe the Future anyway, unless I missed something:
I think a longtermist would say that the effects on future moral people should dominate our moral calculus due to their vast number, not necessarily that they can right now. But we should keep an eye out for how to impact the longrun future positively and take such a chance if we ever see it. Some people think they see the chance now, so they are taking it.* Maybe you will never see something plausible-to-you within your lifetime. But that doesn’t mean a chance will never occur.
For example if we could run amazing simulations to test longrun outcomes, I think a longtermist, if they believed in the tech, would want to give the best-predicted longterm action a go (like, the best sum of experiences had over time, summed at the end), using neutral moral weights for beings living today vs 3 years from now vs 1000 years from now. Of course there will be wider ranges and larger confidence intervals but you’d factor those too, as to what option you’d want to do. By contrast a neartermist would add moral weight to the consequences and experiences for beings existing in the nearterm on top of the differing ranges and confidence intervals which are just sensible to use for both neartermists and longtermists.
*I’ll note that some extinction risks seem low enough percentage chance of happening to me that they may not be worth working on unless you do add neutrally-weighted future generations into the calculus. But it depends on the moral discount rate you’d use, like if your expected rate of population growth is large if everything goes well, and your moral discount rate is gradual enough, you might still end up preferring a “longtermist” intervention, even though you might not really be a true “longtermist” philosophically because you are still claiming that future generations are morally worth less (regardless of confidence), but focusing on the longterm just passed your bar anyway because of the scale.
Longtermism includes the claim that improving the future is tractable. I think that was probably a mistake, and it should just be a claim about values.
Daaaang yeah that seems wrong to me too unfortunately. I can imagine we have passed the threshold on coordination and technology that changing the long-run future somewhat predictably is a tractable cause now (and I can also imagine we haven’t), but I’m pretty sure there were many years in history where it would have been impossible to predict past 5 years ahead. It seems to me that a philosophical position (which longtermism claims to be) should be able to have existed then too. But if it included tractability, that position was likely impossible to hold (correctly) at some moments in history, or in many single-actor thought experiments. But I’m no philosopher
I am unsure if that makes you not a longtermist. My understanding of the basic longtermist claim is as follows:
Future individuals matter as much as those in the present
We should do something to positively shape their lives
I think this has little to do with forecasting accuracy, albeit if we were really good at that, it would help us better outline the something in the second point. I think the major part of it is about how we assign moral value to those who will exist in the future.
I think he’s saying that an important third element is that “We can do something to positively shape their lives”.