You could, in principle, live forever—but if an anvil falls on your head, you will still die.
It is precisely this background risk of accidental death from injury that would limit the hypothetical average lifespan of a non-aging human to roughly around 1,000 years.
With AGI and without extinction or severe permanent disempowerment, digital backups will almost certainly be available long before 1,000 year biological lifespans become relevant. Digital backups don’t need to be about uploading, you could reconstruct biological bodies (including the brain) from the data that specifies them. So anvils shouldn’t be a problem, and distributed backups make catastrophes of arbitrarily large scale fully recoverable, as long as there remains a global civilization that wants the recovery to occur.
And consider this: if humans become capable of fully defeating aging and death, would they not also be capable of modulating brain neurochemistry to adjust levels of plasticity—and therefore openness to new experience? Of course we would.
Apart from any specific issues, there is no a priori reason to expect extremely long term survival of a person on the substrate of a human brain, any more than natural biological immortality for the body. It’s likely a very complicated philosophical and technical problem for people to remain themselves over very long lifespans, let alone while growing up to become much smarter than human brain allows.
The philosophical aspect of this problem is more prominent than for biological immortality (or functionally equivalent uploading), so even non-omnicidal AGIs won’t obviously give a satisfactory resolution to it that can be evaluated by a human in a sufficiently short amount of time to survive as the same person on legacy human brain substrate (uploaded or not), along a single timeline of experience.
Am I correct in understanding that you are suggesting that the human brain will deteriorate even without aging, since it is simply not designed to last for hundreds of years?
I know little about the limits of the brain over centuries or millennia, where exactly it will fail, and how this can be repaired. If you (or others) have a good resource with similar arguments, please let me know about it.
Of course, the easiest thing would be to recreate the brain — organic or silicon — with all its memories. But perhaps the problem of identity is even more complex than the problem of preserving the brain over centuries? I’m not sure.
Maintaining perfect biological health (including for the brain) is a more objective target than maintaining an abstract person implemented by the brain, there is more philosophical difficulty in defining what success means. A healthy brain in a million years might just effectively end up containing someone else, in a way that its original inhabitant wouldn’t endorse on reflection. And there might be insufficient time for that reflection to occur in a single lifetime while the original person is still there and didn’t yet become someone else. AGI-written textbooks on the topic might help, but avoiding undue influence via such textbooks is similarly harder to define than autonomy in thinking on your own.
With AGI and without extinction or severe permanent disempowerment, digital backups will almost certainly be available long before 1,000 year biological lifespans become relevant. Digital backups don’t need to be about uploading, you could reconstruct biological bodies (including the brain) from the data that specifies them. So anvils shouldn’t be a problem, and distributed backups make catastrophes of arbitrarily large scale fully recoverable, as long as there remains a global civilization that wants the recovery to occur.
Apart from any specific issues, there is no a priori reason to expect extremely long term survival of a person on the substrate of a human brain, any more than natural biological immortality for the body. It’s likely a very complicated philosophical and technical problem for people to remain themselves over very long lifespans, let alone while growing up to become much smarter than human brain allows.
The philosophical aspect of this problem is more prominent than for biological immortality (or functionally equivalent uploading), so even non-omnicidal AGIs won’t obviously give a satisfactory resolution to it that can be evaluated by a human in a sufficiently short amount of time to survive as the same person on legacy human brain substrate (uploaded or not), along a single timeline of experience.
Am I correct in understanding that you are suggesting that the human brain will deteriorate even without aging, since it is simply not designed to last for hundreds of years?
I know little about the limits of the brain over centuries or millennia, where exactly it will fail, and how this can be repaired. If you (or others) have a good resource with similar arguments, please let me know about it.
Of course, the easiest thing would be to recreate the brain — organic or silicon — with all its memories. But perhaps the problem of identity is even more complex than the problem of preserving the brain over centuries? I’m not sure.
Maintaining perfect biological health (including for the brain) is a more objective target than maintaining an abstract person implemented by the brain, there is more philosophical difficulty in defining what success means. A healthy brain in a million years might just effectively end up containing someone else, in a way that its original inhabitant wouldn’t endorse on reflection. And there might be insufficient time for that reflection to occur in a single lifetime while the original person is still there and didn’t yet become someone else. AGI-written textbooks on the topic might help, but avoiding undue influence via such textbooks is similarly harder to define than autonomy in thinking on your own.