Just taking this seriously pretty much resolves all these problems.
It makes sense to take high cost/reward, low probability events seriously if the expected utility works out. Examples include reducing existential risk substantially, even at cost to short-term utility (say, in the form of well-being) to increase the probability that we can eventually figure out how to optimally arrange matter and energy and trigger a utilitronium shockwave.
Just taking this seriously pretty much resolves all these problems.
It makes sense to take high cost/reward, low probability events seriously if the expected utility works out. Examples include reducing existential risk substantially, even at cost to short-term utility (say, in the form of well-being) to increase the probability that we can eventually figure out how to optimally arrange matter and energy and trigger a utilitronium shockwave.