[ExI] chemo-preservation and fund raising

Anders Sandberg anders at aleph.se
Mon Jun 16 22:06:56 UTC 2014


Max More <max at maxmore.com> , 16/6/2014 7:06 AM:

It would take much more than two months of economic turmoil to endanger Alcor's patients. First of all, that is not enough to "evaporate the liquid nitrogen". We have run a test and found that a Bigfoot dewar (empty of patients, obviously) did not run dry for something like five months. Because the aluminum pods that house patients are conductive, so long as there are even a few inches of LN2 at the bottom, the temperature even near the top is within something like 30 degC of LN2 temperature -- plenty cold enough. The boil-off rates vary, but I think the minimum would be 3 or 4 months.
 
That doesn't count contributions from the bulk tank, which we would use to refill the dewars. 

Considering that there are at least 7 liquid nitrogen vendors in the Scottsdale/Phoenix area, it would take far more than "economic turmoil" to terminate all deliveries of LN2. It would take at least WWIII. In the meantime, we already know that we could go out and acquire a small liquid nitrogen plant and make our own. (We have a powerful backup generator, which I had installed this year, that could power it.) That would cost about twice as much, which is why we haven't already bought one. 
There is something interesting here to consider: what is the duration distribution of industrial outages and economic turmoil? I would love to see some proper data about outages of industrial chemicals and their time distribution. 
I have data about blackouts, and they are typically power-law distributed. I would a priori expect industrial and supply chain outages to also be power-law distributed: industrial and economical systems can also have cascading failures when loaded heavily, and optimization processes may drive towards power-law behavior.http://www.ece.cmu.edu/cascadingfailures/Criticality-nedicPSCC05.pdf
Now, this means that the probability of a failure lasting longer than X has probability X^-a, where a is some positive constant. So if there is a one-day outage, expect one-month outages with a factor 30^-a less probability and one year outages with 365^-a factor probability. If we guess a between 1 and 2, then month outages have 3-0.1% the chance of the one day outage and year outages 0.2-0.0007% of the chance. Now, throwing in a 1% chance per year of a day disruption (which seems reasonable based on 20th history, containing WW II, the Cuban missile crisis and 911) that gives over a century 63% of some disruption, 0.3-0.1% chance of month disruption and 0.2-0.0007% one year disruption (things get linear when dealing with low probabilities). 
Overall I think Max is right: LN2 outages long enough to mess up Alcor are WWIII class... which of course still doesn't make them unlikely enough for our liking (just look at our office guesses https://flic.kr/p/nYWTJV for catastrophic events - these are somewhat above my 1% per year number).


Anders Sandberg, Future of Humanity Institute Philosophy Faculty of Oxford University
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20140617/1290f978/attachment.html>


More information about the extropy-chat mailing list