There’s nothing more certain than taxes
and death, but in the future, maybe not the latter. We’ve discussed concepts like life extension
before, and last week we discussed mind augmentation. Most people have at least a passing familiarity
with the notion that you might be able to backup or upload your mind to a computer and
in that way live forever, or pretty close to it. In the Civilizations at the End of Time series
we’ve discussed the ways a civilization might outlast the very stars themselves and,
indeed, potentially flourish and prosper on a scale that would make the entire stellar
phase of the Universe seem like a brief prelude. In that regard, we are barely into the first
sentence of that prelude. Humanity has existed for only an eyeblink
of time compared to the Universe as a whole, and it is still quite young. We are about 1% of 1% of the way through the
period of time in which stars will form and die, and the period of time in which humanity
has been around is only about 1% of 1% of that. Recorded human history is even shorter, just
a percent of the time humanity has been around, while the individual human lifespan is about
a percent of that. Only about a hundred generations have passed
since the Roman Empire was at its peak, and the dawn of recorded history lies only about
as far again back from then. For most of that time, people typically did
not die of old age, and even what folks meant by old age included a lot of circumstances
we no longer view as natural causes of a long life. At the age of 37, I have already outlived
the supermajority of all humans who have ever lived, yet barring unexpected illness or accident,
I should have just as much time left to me. When we start talking about human lifespan
extension people often recoil from the idea as rather fantastic, but it is worth remembering
that many of the things that used to kill people regularly in the past have been outright
eliminated or reduced to being exceptionally rare. Those people living when such causes were
normal, indeed, more likely than dying of old age, might have considered their elimination
rather fantastic too. We are just getting to the point where technology
is hinting at ways to extend someone’s life enormously, potentially without any limit,
so contemplation of this topic tends to fall into two chief attitudes. The first is disbelief, while the second is
total acceptance, immortality is possible, and perhaps even folks alive nowadays might
enjoy it, not simply folks living in some distant future. This is our topic for today, because we need
to contemplate some serious impediments to extremely long life that come into play once
simple aging is no longer an ultimate expiration date on your existence. Over a long enough period of time you are
likely to have some sort of accident kill you, or to be murdered. Even if your odds were only one in a million
every year, you’d have a fifty-fifty chance of dying in the next 700,000 years and less
than a 1% chance to make it to five million. Now most folks would shrug at that, that’s
a lifespan on a timeline as long as humanity has been around, more than enough, but it’s
important to keep in mind you need to do way better on your survival odds if you want to
be seeing the End of the World. Typically folks would suggest you probably
want to either get a digital backup of your mind or go entirely digital yourself. Now it’s kind of debatable if you, yourself,
can actually go digital or if you are really just making a copy of yourself, and that’s
an important distinction since if you don’t view that copy as you, you might prefer to
just have a copy that can go live if you die. Many folks debate whether or not such a copy
is you, but whether or not it is, it's not something you can really prove or disprove. Regardless, what matters is whether or not
you think it is. But even that doesn’t necessarily mean you
won’t get backed up if the option is easily available, since a person has reason to want
a copy of themselves for more than personal continuity. Putting it bluntly, most of us have stuff
we’re willing to die for: friends, family, causes, etc. But we also have stuff we’re willing to
live for, or in this case, be resurrected for. Most of us have people or projects we would
not want to see ourselves absent from; I would like to know this channel would go on if I
fell over dead tomorrow, and the most obvious successor to operate it is me. Nobody’s going to finish writing that novel
you’ve been working on or tend that garden you’ve spent years improving, not the way
you would. Most of us have something like this we deeply
care about, and that’s not even including our friends and family, let alone our kids. This isn’t some cliché scifi horror novel
either; that copy of you isn’t going to go home to your spouse and little ones and
turn demonic. It is you, it isn’t an ‘it’. It might be on a TV monitor for a while or
in some sophisticated android till a body is grown, or re-grown. Barring murder, in a civilization that can
backup memories, odds are only massive brain damage that shreds almost everything will
keep you dead, but it would be nice if the tiny little robots repairing your neural connections
had a backup copy to look at as a blueprint. I’ve heard folks suggest they’d be freaked
out by a copy of a dead family member, and that’s an entirely legitimate response,
but that’s because we’re not used to it. We’re very good at believing what we want
to be true, and by default we suspect such a thing only because it’s too good to be
true and because we’ve seen a lot of horror movies about bringing back the dead and it
going all wrong. We’ve got a lot more reasons to want to
believe it’s really them or close enough, both from a personal desire and a strictly
scientific perspective, so I’d rather imagine most folks will increasingly tilt to regard
them as a true copy or even just the original who was away for a bit in the hospital. So there’s a lot of reasons to have such
a backup around and not many reasons not to. Lots of potential problems too. Fortunately, one that isn’t too big a problem
is storage space. While the human mind is still a more potent
processor than our best supercomputers, even if that gap is almost closed, human memory
storage is estimated to be in the area of 10 terabytes to perhaps 10 petabytes, and
you can buy 10 Terabyte hard drives these days. A modest gain in hard drive capacity and cost
will permit affordable storage of data equal to even that higher end estimate of human
memory. Right now you could buy the lower end value
for a few hundred bucks, and the higher end for around a quarter of a million dollars. We never want to fall into the trap of thinking
that Moore’s Law and its parallels for other aspects of computing like memory is an actual
law. There are no guarantees we will get computers
even one penny cheaper per unit of memory or processors than we have right now, but
I think we can assume that by the time we have brain scanning technology we will have
at least another order of magnitude or two knocked off the price per byte of memory. Such being the case, the actual storage cost
for a copy of a human mind should be affordable even if it is that higher end, 10 petabyte
value, and of course 10 terabytes already is. What this means is that we have every good
reason to believe we could store a human mind on something smaller than a human brain and
a good deal cheaper than most annual life insurance premiums. A good analogy perhaps too, since a backup
is essentially a type of life insurance. You get those to take care of your loved ones
after you’re gone, a backup of you does the job better. But it also means you can probably have a
ton of backups, especially if memory gets a lot cheaper. Most likely the cost bottleneck would be all
about the transmissions themselves, and of course the scanning equipment. In a realistic scenario, your default 22nd-century
individual, who is a little bit of a cyborg, probably has a mind–machine interface that
can double as a brain scanner. Let’s use a specific case, we’ll say Steve
from Austin, Texas. Steve has neural lace woven around in his
brain and scanners in his skull and little nanomachines that help to monitor and repair
his brain. They tend to do most of their work at night
and he doesn’t use his bandwidth much then so they assemble a snapshot of his brain and
transmit it to a hard drive implanted into his leg bone with a fiber optic running down
to it. That’s his first backup, though really there’s
one already in his head, but it’s ever shifting and not much of a backup since it would likely
be damaged by whatever damaged his brain. Steve’s pretty tough and those little nanobots
could probably repair his brain on their own if he got stabbed in the skull, probably without
even having to check that backup black box in his leg. He’s also not really worried about losing
a day’s worth of memory, so he only updates his backup at night, once a day. But he is transmitting all the time, like
most people he has a deadman switch constantly monitoring his lifesigns and recording sound
and video near him for his secure storage, that will be transmitted to a security company
and then the police if anything goes wrong. He can always watch those recordings if he
loses that last day of memory too. Whatever amount of bandwidth you might need
for transmitting a backup is way more than a few megabytes of video, audio, and biometric
data. Of course when Steve sends his backup daily,
to a local datahub or warehouse, he doesn’t send a whole copy of his brain, even if that
is only 10 Terabytes that’s still a lot of data to send down the wire or wirelessly. He’s probably just transmitting a comparison
of his current mind with his last copy and filing those changes. All the new neuron connections and changes
in potential and such. Now Steve is a bit paranoid, so from that
datahub the data gets sent to two different companies, each of which keeps its own redundant
backups too. If you are in the brain-copying business,
you want to make sure you’ve got a reputation for having giant armored vaults with backups
and the kind of security that makes Fort Knox look like a convenience store. He’s got two backups against those companies
failing. It’s like with getting yourself frozen,
folks ask what you should invest your money in to make sure you still have some when you
get thawed out, to which the answer is, the company that froze you. If they are successfully thawing people out,
their stock is probably looking pretty hot, and if their stock isn’t too hot, you probably
ain’t staying too cold, so the status of your investments is the last thing on your
mind, or what’s left of it. Multiple redundancies, prior copies that haven’t
been updated for a while just in case someone has been tampering with your uploads or even
hijacked your implants to send static or gibberish so they could kill you later when all those
stored copies are gibberish too. Encryptions on those transmissions, regular
checks with the technicians to make sure all is well. This makes somebody very hard to kill, for
keeps. They can still get your body of course, one
immediate application of this technology is that you can copy your mind into an android
and go shoot someone with a death ray that disintegrates them. That would probably be a pretty common form
of interplanetary tourism too, since it’s faster and probably cheaper to get your brain
scanned over to an android on another planet for a while. You might get some interesting companies that
provided anonymous brain scan storage or don’t ask don’t tell android rental. Though I think you could still track them. But if you’ve got that tech you also have
off planet storage too, just having androids sophisticated enough to handle a human brain
copied into them grants you way easier off-planet construction abilities, but that’s not too
relevant, since all the tech needed for such things allows really good automation and self-replication
anyway. So Steve’s other copies are off at the Hyperion
Data Repository on Titan, and with a small city-state in the asteroid belt that is the
Swiss Bank of the Brain Biz. Those only get copied once a week, but they
keep 100 prior versions before copying over and Steve assumes that he only needs those
if either Earth is getting blown up or someone has engaged in an elaborate murder conspiracy
against him. At least that’s what Steve’s wife Jaime
thinks. Steve is pretty paranoid so he actually has
a third copy she doesn’t know about that he only transmits once a month and never from
home. After all, no encryption is ever safe from
someone with the password or access to the account a password reset goes to. Steve heard of a man killed by his wife after
they had a very bad argument, where she didn’t want him dead, just wanted his memory of the
argument erased. Steve’s fairly normal as folks go, in terms
of paranoia and backups, he’s not a secret agent or a test pilot or something dangerous
that might get his body incinerated. But he’s pretty hard to kill, and indeed
probably a lot less likely to die, permanently, than one in a million per year. It can pretty much only happen to him if he
is murdered, and we already have a fair few places where the murder rate is nearly that
low. As we said at those odds your half-life was
about 700,000 years. So he might live quite a long time, with even
lower odds. This is even bigger for strictly digital entities,
especially ones that don’t feel a need for speed. As we’ve mentioned many a time before, the
signals running around your brain generally move slower than the speed of sound, often
a lot slower, and it’s only a millionth the speed of light. If you copied a human brain exactly, but spaced
out to run at light speed but at the human rate of consciousness, that brain would be
the size of a planet. Now you’d probably have that bundled into
all sorts of nodes but it means someone can spread their intelligence all over the planet
and still maintain the normal human rate of thought without needing a single extra bit
of total processing power. That’s a very hard target to take out, when
they might have a dozen redundancies for every node and thousands or even millions of such
nodes. This doesn’t even include actual backup
copies of them elsewhere or even running clones. There’s a million asteroids in the Belt
big enough to make comfortable city-states for regular humans, and they could easily
have tons more entirely digital people running around in the background, potentially with
their consciousness distributed over several light seconds. And if you don’t mind going slower, or adding
more processing, you could potentially distribute your awareness like a gas cloud spread over
entire solar systems or other solar systems. If your big goal is to survive as long as
the stars or longer, you might not mind if a thought took a year, if it protected you
from pretty much any conceivable attack. Helps with boredom too, one of those big threats
to continued existence. That’s the big one of course. How do you kill an immortal? In a fantasy novel, they’ve presumably got
some special weakness, ya know, throw their precious ring into a volcano, or there’s
a magic sword that can kill them. We don’t get those in the real world, but
hypothetically, you could get them with some virus that corrupted them and their copies,
they presumably need access to those. But they might have disconnected backups or
ones that were read only in someone else’s vault as a last fallback. You could go get all those storage repositories,
but you need to know where they all are. You need to be able to blow them up. You need to be able to deal with the issue
that other people are probably on those too, and you need to get them all at once or inside
the time window light lag provides or they’ll just copy elsewhere. And all the security for this is to protect
human lives, many human lives, so odds are you are going up against tons of the best
minds around who designed those security systems. Barring such options, how can you kill an
immortal? The easiest way is to get them to kill themselves. There’s more to that than simple suicide,
but we can start there. People probably do not want copies of their
minds floating around that they don’t know about or don’t have access to. That is your privacy after all, so having
some backup you’ve removed from your own memories triggered to go off if you committed
suicide would seem a special level of paranoia, and you can’t update it all the time or
you will need to know about it or otherwise leave evidence it exists, so you could be
losing years or decades of memory. Excluding that as probably uncommon and impractical,
a person should be able to kill themselves, and it’s unlikely any court would rule that
any decision you made in the past, like a clause preventing you from wiping your backups,
was something you couldn’t invalidate. Maybe though, you might have some weird cases
of divergent multiple persons with the same original mind where one tried to delete the
original backups and got refused. Now it sounds kind of absurd that you could
talk someone into killing themselves without overt mind control, but two notes on that. First, while I always say I can’t imagine
ever getting bored with life to the point of wanting to end it, that is the most common
rebuttal to life extension I tend to hear when we discuss the topic. Folks would get bored and choose to die. I, of course, always say that’s their own
business and doesn’t have any relevance to whether or not such tech should be developed
or if other people should be allowed to use it. I also don’t think most people would get
bored either, but I could be wrong. Add to that, we expect psychology to keep
improving especially when we are at the brain scan and emulation level. A trained expert working on someone who was
already kind of bored might be able to talk them into ending themselves. They almost have to have access to every recent
backup at least of their mind, so they can delete those. You could have things in place to prevent
suicide, if you were worried that you might be struck by extreme but temporary depression
at some point in a multi-million year life. But you probably don’t want strong anti-suicide
restraints in place either, considering you could end up being truly immortal to the point
you can’t even kill yourself. Your core architecture prevents you from even
trying. That’s how you end up a trillion-year old
miserable entity trapped for all eternity, potentially even driven to steal resources
from others to keep going after the stars are gone because your internal mental monitors
regard that as akin to starving yourself to death and are allowed to force you to eat. It’s a little chilling to imagine a dark
and cooling Universe populated only by resource raiders cannibalizing each other, and in some
ways that seems even worse if each of them is secretly grateful when they lose and can
finally die. So while you might want to build in time constraints
on deleting backups, waiting periods to confirm the request and so forth, you still probably
want that option. If it is there that’s one way to get an
immortal, persuading them to take a nihilistic view of life or even just convincing them
that life will be exciting again with real dangers and few to no backups. Of course at a fundamental level, if your
psychology is getting that good, and so is your neuroscience, you might be able to ‘kill’
them just by changing them. Fundamental life changes can have a big effect
on a person, and permeate out to change their attitudes on all sorts of things. And that’s without precision psychology
or the ability to simply delete a given desire. For a digital entity, one can presumably be
sufficiently in control to flip a switch that makes you like chocolate and hate strawberries,
where before you didn’t like chocolate and loved those berries. You don’t have to move fast either, revenge
is a dish best served cold, and you are presumably just as immortal as they are. If it takes you a million years of subtle
manipulation to effectively kill off the old persona with one so different it doesn’t
act the least bit the same, what do you care? Which brings us to our last point. You can use all these methods to extend your
life indefinitely, but how much is that person really you? Yes, the changes are gradual, but does that
really matter? In general, we also have this same concern
with Transhumanism. If you boost the mind up to superhuman levels,
is what remains actually you or did that person die? Not a seed the tree grew out of but just the
decaying logs of a previous trees it used for nutrients and root structure? This is one reason I can imagine for why people
might skip on mental augmentation or super long lives. They just don’t believe that final product
would be them, because it's beyond simple gradual change so that the final person is
no more them then we are the various dirt and microorganisms that were around when the
dinosaurs roamed the planet. Now you could put safeguards in to prevent
such major changes, to make sure you never drifted outside the acceptable ranges of thought
and behavior you wanted. Have certain things hardwired as it were. Or you might use resets, effectively being
like Leonard Shelby from Christopher Nolan’s Memento, the reverse of the groundhog day
where you keep repeating the same day, but remember it; here you keep experiencing new
days, but forget the previous ones. Something like that could happen just from
not being able to contain all your memories, not just running out of space for them but
for the indexing system you use to recall a billion years’ worth of life. Faced with options like that, a lot of folks
might decide enough is enough and opt to shut down, not from boredom, but from recognition
that they just weren’t the same person anymore, and either they archive all those old memories
entirely outside their new mind or just end it entirely. Either way that original person is gone. That’s a fairly dismal outlook on eternity,
but this is our Halloween episode after all. There could be solutions too, we’ve hardly
explored the options in detail for this problem we can’t experience yet anyway, and figuring
those out will give future generations something to do so they don’t get bored. For mind uploading, I would expect it to be
a pretty major sector of interplanetary trade, our topic for next week, simply because if
it becomes normal, most folks would want a backup far from where they lived. It will also play a big role in how people
approach warfare in the future too, when we revisit that topic early next year. For alerts when that and other episodes come
out, make sure to subscribe to the channel, and if you enjoyed this episode make sure
to hit the like button and share it with others. And join in the conversation in the comments
below or at our facebook group, Science and Futurism with Isaac Arthur. Until next time, thanks for watching, and
have a great week!
