Actually I'm against the combination
Ineffective altruism is good for signalling without breaking the correct assignment of credit that the world needs.
It becomes a problem when it's effective though.
Altruism should (mostly) be kept ineffective
#effectivealtruism
Ineffective altruism is good for signalling without breaking the correct assignment of credit that the world needs.
It becomes a problem when it's effective though.
Altruism should (mostly) be kept ineffective
#effectivealtruism
If you want the ultimate powerful take, it's that you shouldn't even contribute to existential risk mitigation charities, because (in a large multiverse) it's better if our civilization is selected out so that most future experience moments come from civilizations that are better
The longtermist case against existential risk mitigation charity in big worlds
Some people are asking how you should know whether your civ is "better than average" or not.
This is not easy to answer. So, much uncertainty here.
Also, even in the face of uncertainty this strongly favors accellerationism.
This is not easy to answer. So, much uncertainty here.
Also, even in the face of uncertainty this strongly favors accellerationism.
But, of course, that's conditional on big worlds being true.
If we live in a small world, then this reasoning does not apply.
If we live in a small world, then this reasoning does not apply.
A plausible basilisk:
If you notice that your world is obviously morally broken, it matters a lot that it doesn't make it past the singularity, and worlds that re not broken may acausally reward you for "aborting" broken worlds, to raise the average across the multiverse.
If you notice that your world is obviously morally broken, it matters a lot that it doesn't make it past the singularity, and worlds that re not broken may acausally reward you for "aborting" broken worlds, to raise the average across the multiverse.
Still, it's hard to really tell where the cutoff is. And most of the benefit may be in risky accelerationism where success is correlated with things going well
i.e.
99% chance of extinction, 1% chance of great future >>
50% chance of extinction, 50% chance of hell-universe
i.e.
99% chance of extinction, 1% chance of great future >>
50% chance of extinction, 50% chance of hell-universe
Hence e/acc may actually be good
Like, if you're doing e/acc, you are implicitly on this train (high risk high reward)
People asking for a "long reflection" to minimize x-risk and "solve morality" may be doing exactly the wrong thing - such a period is likely going to be used by moral parasites to thoroughly corrupt civilization and result in a hellworld, which we do not want to amplify via AI.
So maybe the correct policy, given big worlds, is R-selection for the singularity (small chance of big success), rather than a failed attempt at K-selection.
Loading suggestions...