I do have one useful comment. Arguments like this are used to justify no attempt at the slightest life extension now. “since even if I could survive today, I might die at some far future date, perhaps when I get bored of life in 10k years or when the universe runs out of fuel in a few trillon years. So I’ll take oblivion right now, at a mere 75 or 85 years because in the end what’s the difference”
This argument seems flawed because quantities matter. Comparable to allowing a 5 year old to commit suicide because they are bored with life and want to make room for another 5 year old.
And second, while there might or might not be a solution to the heat death of the universe, there is a solution to your ’world is too crowded with immortals, kill them”. If you absolutely wanted to reduce crowding but were against death, merges are a viable solution.
Any technology that can fix every possible flaw a human body could encounter and keep their mind running through any insult or injury by it’s nature has the ability to capture, back up, store, and even merge human memories. A merge could be as simple as finding another immortal human with very similar personalities and life experiences and de-duplicating them. For example the trivial example is that if you made a clone of yourself yesterday, and both of you have only experienced today as a divergence, a merge would cause you both to remember today (and you would be reduced to a single merged-self running on one set of hardware). This is actually easier than it sounds if you think of memory as a simple set, where now you have 2 partially overlapping keys when you think of “today”. Just now you have 2 keys for today, one from yourself, one from your clone.
We can trivially do such merges with AI agents using neural networks today, this is how OpenAI does training in parallel over large clusters.
Some memories would become ambiguous “what did you do today”, but you would gain the skills update from both you and your clone self.
Anyways you can merge as often as needed—a little data gets lost in each merge—but you can use merges to reduce the population until it can support however many new individuals you believe is ethical.
Merges are a general solution to most problems you would encounter with a technology that allows for copying of memories.
Yes I definitely agree that this is not an argument against research on life extension.
Merging is very interesting, it seems that it would also prevent AI’s or humans who are self-copying too much. Agents could also merge in a temporal sense by say, alternating the days that they run, or by running more slowly (and thus using less resources).
I do have one useful comment. Arguments like this are used to justify no attempt at the slightest life extension now. “since even if I could survive today, I might die at some far future date, perhaps when I get bored of life in 10k years or when the universe runs out of fuel in a few trillon years. So I’ll take oblivion right now, at a mere 75 or 85 years because in the end what’s the difference”
This argument seems flawed because quantities matter. Comparable to allowing a 5 year old to commit suicide because they are bored with life and want to make room for another 5 year old.
And second, while there might or might not be a solution to the heat death of the universe, there is a solution to your ’world is too crowded with immortals, kill them”. If you absolutely wanted to reduce crowding but were against death, merges are a viable solution.
Any technology that can fix every possible flaw a human body could encounter and keep their mind running through any insult or injury by it’s nature has the ability to capture, back up, store, and even merge human memories. A merge could be as simple as finding another immortal human with very similar personalities and life experiences and de-duplicating them. For example the trivial example is that if you made a clone of yourself yesterday, and both of you have only experienced today as a divergence, a merge would cause you both to remember today (and you would be reduced to a single merged-self running on one set of hardware). This is actually easier than it sounds if you think of memory as a simple set, where now you have 2 partially overlapping keys when you think of “today”. Just now you have 2 keys for today, one from yourself, one from your clone.
We can trivially do such merges with AI agents using neural networks today, this is how OpenAI does training in parallel over large clusters.
Some memories would become ambiguous “what did you do today”, but you would gain the skills update from both you and your clone self.
Anyways you can merge as often as needed—a little data gets lost in each merge—but you can use merges to reduce the population until it can support however many new individuals you believe is ethical.
Merges are a general solution to most problems you would encounter with a technology that allows for copying of memories.
Yes I definitely agree that this is not an argument against research on life extension.
Merging is very interesting, it seems that it would also prevent AI’s or humans who are self-copying too much. Agents could also merge in a temporal sense by say, alternating the days that they run, or by running more slowly (and thus using less resources).