Calculating the probability of immortality

How likely is it that a given object will survive forever?

With many groups predicting that human immortality is just around the corner, you could say we all have a vested interest in the answer.

A t first glance, the odds are not good. As David Eubanks of Coker College in South Carolina puts it:

“Imagine that some subject survives each year (or other time period) with a probability p. Assuming for a moment that p exists and is constant over time, it’s easy to compute the dismal odds of long term survival as a decaying exponential. Unless p = 1, the probability of n-year survival approaches zero.”

In other words, the probability of surviving forever is exactly zero.

But this suggests a strategy: the route to immortality is to find a way to increase this probability over time.

Suppose the object we want to make immortal is the data on a hard drive. Then copying the data to another hard drive each year should ensure that after n years there are n copies. If one drive fails, we can easily reconstruct the data onto another drive. So unless all the drives fails at once, the data should be immortal.

Except for one problem. That approach ignores global catastrophes such as comet strikes which would destroy all the drives in one go.

Eubanks says there are two ways to tackle this problem. Life has found one of them which is to produce many diverse copies of the same thing, spread them around the planet and make them work in different ways ie exploit different energy sources.

The other is for a single organism to use its intelligence to avoid catastrophe. “It must collect information about the environment safely and inductively predict it well enough to avoid death,” says Eubanks.

How do these two strategies compare? Naturally, single individuals have a harder time of it because it’s tough to predict and adapt to all possible catastrophes. As Eubanks says, “We’re the product of billions of years of creatures that survived long enough to reproduce, and therefore have very deep survival instincts. And yet we can fall asleep while driving a car.”

Eubanks is even more pessimistic. He says that simulating every possible environmental disaster may be tricky enough but such a disaster would then force us to re-evaluate the way we evaluate disasters and so on ad infinitum. In short, it’s a calculation we are extremely unlikely to ever be able to undertake, making it hard to think of a way we could improve our probability of survival, year in year out.
The bottom line is that humans are unlikely to survive forever and neither is intelligent life anywhere else in the Universe.

“This speaks to the Fermi Paradox, which asks why the galaxy isn’t crawling with intelligent life.”

Quite. But Eubanks’ paper has another sting in the tail.

“The conclusions of this paper could lead one to believe that a democratic government cannot focus solely on external threats, but should also be constantly trying to improve the chances that it does not “self-destruct” into tyranny.”

What democracy could he be thinking of?

Ref: arxiv.org/abs/0812.0644: Survival Strategies

One Response to “Calculating the probability of immortality”

  1. Here are a couple of the more fanciful notions I left out of the paper:

    1) What if the Intelligent System self-modifies to exponentially increase its subjective time? Then it can “live forever” in finite classical time. I wrote a fanciful short story about two computer processes (which run at hyper-speed) arguing about whether biological creatures could have ever been considered intelligent, since their thought processes were so slow. The story is here: http://zzascape.blogspot.com/2008/04/interruption.html

    2) What if an advanced civilization created seed copies of itself (worlds like ours, for example), and a means to retrieve information to ‘reboot’ the mother civilization through some galactic internet. These reboots would be going on constantly with various copies in different stages of completion on different planets. But they would only be actually needed if the mother civilization had been destroyed, so these seeds would constantly have to be reset (probably destroyed). This is basically a hardware/software delineation. It would make me a little less happy about finding a SETI signal telling us how to improve ourselves. 😉

    -dave