Tikalon Blog by Dev Gualtieri

Survival of the Smartest

April 17, 2013

Creative thoughts often come to individuals while they are in a semi-sleeping state, as when just going to sleep , daydreaming , or just spacing-out from sheer boredom . A recent book,[1] "Finding The Next Steve Jobs: How to Find, Hire, Keep and Nurture Creative Talent," coauthored by video game pioneer, Nolan Bushnell , recalls one innovation that Steve Jobs initiated while working at Atari . He allowed employees to take short naps during working hours to stimulate creativity when they were awake.

I was in space-out mode in 1980 on a taxi trip from the airport to a conference hotel in Dallas, Texas . The cab stopped at a traffic light in back of a package delivery truck (a brown truck, so you'll guess the company). I noticed a number on the back of the truck, which appeared to be the serial number of the truck.

I started to think about how many such trucks there might be, based on this single observation. A crazy thought? Well, physicists have a way of making sense out of crazy things. There's quantum mechanics as one example. In fields such as particle physics , where your data can be understood only by statistical inference , the guiding principle is the confidence interval .

When hunting for things like the Higgs Boson , particle physicists require their statistics to be at the five sigma level.[2] This means there's slightly better than a million to one chance that nature had just organized her random variations to make it appear that there really is a Higgs Boson when none exists. Five sigma gives us 99.99994% confidence that what we've seen is really there.

The normal distribution in which μ is the mean , and σ is the standard deviation . Most of the area under the curve falls in the two-sigma interval between ±2σ. (Via Wikimedia Commons.)

In most other cases, scientists don't demand such accuracy. This is because our experiments aren't as controlled as those on the Large Hadron Collider , and we would never be able to publish any results if everything needed to be five sigma. Instead, we replace the gold standard of five sigma by a lesser standard of just two sigma. Two sigma is 95.5% confidence, although we usually just refer to it as the 95% confidence level.

What could I statistically infer from the serial number on that truck? If I require just 95% confidence in my answer, I can claim that this serial number might be at the point 2.5% into a long series of numbers ; or, it's at the 97.5% point. That's a big difference, but it does give me an idea of how large, or huge, this package delivery company may be. If the number I saw was 10,000, the statistical inference is that this company could have somewhere between (10,000/0.975) and (10,000/0.025) trucks; that is, 10,256-400,000 trucks.

I wrote about such statistics in an article about the Doomsday argument (The End is Nigh (Maybe), July 27, 2007 ). The Doomsday argument has been around for quite a while, the most recent formulation being the one of Princeton University astrophysicist , Richard Gott.[3] As its name implies, the Doomsday argument is about the endpoint of the human race , our extinction .

Gott doesn't cite any specific extinction factors, although there are many from which to choose. Instead, he considers extinction to be a random occurrence, with statistical significance in the fact that we're still here. Gott uses the 95% confidence level to calculate the range of values for the total number of humans, much like I did for the package delivery trucks. Based on the number of humans who have lived to date and a few assumptions, Gott predicted that humans will persist for another 0.2 million to 8 million years, which is not very long on a geological time scale .[3]

Sir Martin Rees, Baron Rees of Ludlow, Astronomer Royal since 1995, and President of the Royal Society between 2005 and 2010, published a book in 2004 that's even more pessimistic.[4] Rees doesn't think the random argument is plausible, and he lists quite a few species -ending scenarios, as follow:

• Impact by an asteroid or comet .
• A particle accelerator might create an Earth-swallowing black hole .
• Global warming .
• Worldwide pandemic , perhaps caused by a virus created by a terrorist organization.
• Artificial intelligence is created, and we reach technological singularity , when humans become irrelavant.
• A gamma ray burst .
• Nuclear holocaust .
• Overpopulation ; although there's a feedback mechanism, famine , that limits population.

Asteroid Impact at K-T extinction (NASA)

The impact of a huge asteroid is thought to be the reason for the extinction event at the boundary of the Cretaceous and Tertiary periods, 65 million years ago.

Geology changes slowly, but scientific terminology changes much faster. The Tertiary Period is now divided into the Paleogene and Neogene Periods, adding to overall confusion for many students.

(NASA illustration by Don Davis, via Wikimedia Commons.)

The only item on this list not potentially preventable with today's technology is the gamma ray burst. NASA , for example has been taking asteroid tracking very seriously. We should start debating whether we should stop the technological singularity. Taking a plot line from Kurt Vonnegut's novel, "The Sirens of Titan ," this singularity might be humanity's purpose .

A more optimistic scenario for human existence has just been published on arXiv by scientists from Oxford University (Oxford, UK) and Tufts University (Medford, MA).[5] Instead of just focusing on humans, this study considered the fraction of civilizations in the universe which are long-lived. The "universal doomsday " argument argues that there should be very few since, we, ourselves, are not one of them. The study did a careful analysis of the universal doomsday argument, and the result was that longevity of a civilization is possible if the number of early existential threats is small.[5-6]

Longevity is a relative term, since the study predicts just a small fractional chance (a few percent) that we will become a galaxy-colonizing civilization. The key is to devote technology to allay threats such as those listed by Rees, and to actually colonize the galaxy. A distributed population has a better chance for survival of at least some members.

Probability for survival of civilization

The survival probability is graphed as a function of the ratio R = N_L/N_S, where N_L and N_S are the populations of long- and short-surviving civilization. The s values correspond to the number of threats (see reference).[5]

(Fig. 6 of ref. 5, via ArXiv.)

Survival of the Smartest

References: