CRYONICS ANDTHE SINGULARITYBy James D. Miller

(The technological advancesthat thry willlike!J give us in

areas such as nanotech will (ifAI doesn't give it to us first)

significant!J progress the dqy rywhich Alcor will be able to revive

its cryonics clients. "

•

Icame to Akor through contemplatingthe Singularity.Previously, I had heardof cryonics but didn't have enough

trust in human innovation to believe that afrozen body could ever be revived. What Ididn't take into account was the likelihoodthat artificial intelligence (AI) would soonsupercharge scientific progress.

Imagine that some biologist laughs atyour cryonics membership, claiming thatthe technology to bring back someone pre-served in liquid nitrogen is unimaginablyfar off. You might argue with the biologist'spremise, citing steady advances in nano-technology. Alternatively,you could ask thebiologist if he thinks that a million yearsof study would be enough time for scienceto crack the cryonics restoration conun-drum. If the biologist, confident that anaffirmative answer wouldn't give credenceto cryonics, does admit that a million yearswould suffice, then next ask him whetherhe thinks that within eighty years it will bepossible to create a computer as skilled atconducting research as a human scientistis. If the biologist answers yes to this finalquestion then you've practically won yourargument.

Because of Moore's Law, the amountof computing speed you can buy per dollarbasically doubles every year and softwareadvances further this long-term trend ofexponentially increasing computer perfor-mance. Let's say that within eighty years wedo create a human-level AI. If the softwareadjusted computing power doubles everyyear after this it would take only twenty

more years for a singleAI to have a milliontimes the power/speed of a human sci-entist. Even if we soon hit an upper limiton the speed of computers, as long as theprice per dollar of computing power con-tinues to fall at an exponential rate, not atremendous amount of time will transpirebetween when a human-level AI arises andwhen Alcor could afford to hire the com-putational equivalent of a million of theseAIs to research cryonics.

Even without AIs, increases in hu-man intelligences will soon almost certainlyaccelerate technological progress. It's ex-traordinarily unlikely that evolution justhappened to stumble upon the smartestpeople our phenotype can support when itcreated individuals such as Albert Einsteinand John von Neumann. The exponen-tiallydecreasing costs of gene scanning willsoon allow us to find the genetic basis ofgenius. (The Chinese are already looking!)And it won't be too long after finding thisthat someone (such as the pro-eugenicsChinese) will create people vastly smarterthan have ever existed. The technologicaladvances that they will likelygiveus in areassuch as nanotech will (if AI doesn't give itto us first) significantlyprogress the day bywhich Alcor will be able to revive its cryon-ics clients.

Of course, enhanc~d intelligences willdo a lot more than make cryonics practical.They would give us what's call a singulari-ry--a threshold of time at which AIs thatare at least as smart as humans, and/ or aug-mented human intelligences, radically re-

www.alcor.org Cryonics/July-August2012 21

make civilization. Singularity expectationsshould be a game changer for the cryonicsindustry.

Because a successful singularity wouldmassively increase the wealth of mankindit would reduce the resources Alcor wouldneed to have saved to revive its clients. If,as many in the Singularity community bs-lieve, the Singularity comes about throughan ultra-AI as above us in intelligence as weare from ants, then if this ultra-AI valueshuman life it would revive all of Alcor's pa-tients at a trivial cost to itself.

Singularity expectations eviscerateseveral anti-cryonics arguments. I've heardsome object to cryonics because theywouldn't want to wake up in a world inwhich they had no skills. Well, after a goodsingularity we would all be so rich that thiswouldn't be impoverishing, and if the Sin-gularity arrives through ultra-AI then ev-ery human (not just the recently revived)would either have easily upgradeable skillsbecause of their computer implants or behopelessly obsolete compared to what su-per-intelligent AIs could accomplish.

The Singularity eliminates the moralcomplaint that cryonics is unjust becausethe earth is already too populated. As aneconomist I reject this overpopulation ar-gument because I recognize that people arean extremely valuable resource. As a Sin-gularitarian I further reject it because welive on a tiny planet in what appears to bean empty galaxy and the Singularity wouldmake space travel easy.

Singularity expectations should miti-gate concerns that you might be restoredin a world in which all of your friends andloved ones have already died. Leading sin-gularity scholar Ray Kurzweil has pegged2045 as the most likely date for a kind ofsingularity that would almost certainly yieldthe technologies needed for cryonics re-vival. 2045 is close enough that most of uscould expect to have some family membersstill around.

Singularity expectations open up a stor-age option for Alcor. If you consider your-self to have existed ten years ago then youmust be defining yourself by the informa-tion contents of your brain rather than theexact physical components of your body,since all the atoms in your body changed

over the last decade. Consequently, if Alcorcould create a digital copy of your brain atsufficiently high resolution it should be pos-sible for a future ultra-intelligence to recre-ate your brain and bring you (yes, the realyou) back. Unlike a physical brain, a digitalcopy put on a sturdy hard drive could becheaply and indefinitely stored. Alcor couldprobably put high quality hard drives in aplastic bag, bury them in the Arizona desertand expect them to be readable thousandsof years later.

Ideally, Alcor would create this digitalcopy while maintaining your physical brain.Alas, the best way to scan a brain is to firstcut it into extremely thin slices. Our currentscanning technology isn't good enough tocapture most of the details of your brain,so if you have trust in Alcor's ability tomaintain your physical brain you shouldn'twant them to destroy it to get a digital copy.Still I propose that Alcor ready a slicing andscanning doomsday plan against the con-tingency that it can't maintain cold storagebecause of, say, financial problems, gov-ernment intrusion or even the collapse ofcivilization. Alcor could also offer an ex-tremely cheap "cryonics" option where allit does is create and store this digital copy.As scanning technology improves it wouldbecome more and more likely that a post-Singularity super-intelligence could bringyou back from the digital copy.

At some point it might become pos-sible for Alcor to make extremely highresolution but non-destructive scans ofits patients' brains. Copies obtained fromthese scans would provide some insuranceagainst accidents.

The most actionable implicationof singularity expectations is that Alcorshould advertise in the Singularity commu-nity. Already many well-respected membersof this community have signed up for cry-onics, including:

Ray Kurzweil, inventor and leadingSingularity intellectual

Robin Hanson, economist

Eliezer Yudkowsky, leading friendly-AI theorist

Michael Anissimov, Media Directorfor the Singularity Institute.

The Singularity is a variance boosterthat might end up destroying everything wecare about. But if the singularity does gowell it would deliver a utopia so fantastic asto be beyond anything we can imagine. AI-cor could pitch itself with the slogan "don'tdie and miss out on the Singularity."

In my book, Singularity Rising: Surviv-ing and Thriving in a Smarter, Richer and MoreDangerous World, I end the introductorychapter with the following paragraph:

This book has one recommenda-tion that, if you follow it, couldradically improve your life. It's aconcrete, actionable recommenda-tion, not something like "Seek har-mony through becoming one withCreation." But the recommenda-tion is so shocking, so seeminglyabsurd, that if I tell you now with-out giving you sufficient back-ground, you might stop reading.

Guess what I'm referring to .•

James D. Milleris an associateprofessor ofeconomics atSmith College and

. was a speaker"'~. at the 2008

Singularity Summit. He is the authorof Singularity Rising, to be publishedon October 2, 2012, and CameTheory at Work. He has a Ph.D. ineconomics from the University ofChicago and a J.D. from Stanford.He joined Alcor in 2008.

Endnotes:Author's Homepage

www.JamesDMiller.orgSingularity Institute

http://singinst.orglBeijing Cenomics Institute

https://www.cog-genomics.orgl

22 Cryonics/July-August 2012 www.alcor.org