May 10, 2019
To be posted on Nobabies.net
Uruguayan Scientific High Performance Computing Initiative
University of the Republic
Av. 18 de Julio 1824-1850, 11200 Montevideo, Uruguay
Dear Sergio Nesmachnow,
You are quoted in “Scientific Computing with Limited Resources,” Nature vol. 569 no. 7754 May, 2, 2019 page 147 and among the experts presenting only you mentioned an amount of money. $10,000 is a little out of my reach; I think I could put it together in a couple of years, but at my age I am not sure to live so long. $5,000 I could manage now for something that might contribute to my life’s work, and what I have to offer is mission critical. It involves some calculations that need about 100 gigs of RAM. My desktop has 32 gigs and could be upgraded for another thousand, but I am limited by energy appropriate to my age. When I wrote the program, I was younger, and even then, it was a miserable experience. I need you and you do not know how much you need me, so it falls to me to trust you.
The fact is that fertility in humans out past a few generations requires generally marrying cousins, about 7th or closer. If outbreeding is the general rule, birth rate will soon fall below replacement and a few generations go to zero. Rich countries like the US and Uruguay have turned the corner with total fertility below replacement and age at first marriage rising inexorably. The enclosed DVD gives the logic of why this had to be, the empirical proof that it happens in humans and other animals, a glimpse into the predictable effect on history and laboratory evidence that the effect is mediated by a combination of pre-and post-zygotic mechanisms caused by mismatch of methylation patterns in certain stretches of DNA. I cannot think of any scientific law supported by so many lines of hard data before becoming universally recognized. But in spite of thirty years of dedicated work I have had no success in warning people. Yes, I’m sure there is a political dimension I do not understand, but I do believe that people will make more prudent choices if given the truth.
I have written a computer model of my theory about how it works, and the program successfully models a number of population trajectories in humans, mice, fruit flies and voles. But it is very noisy. The present form of the program only needs about a megabyte of RAM. To my eye, if noise were reduced by a factor of 10, it would be far more convincing. That, I suppose, means where I have a hundred sites that can by methylated (allowing multiple methylations per site) it should rise to 10,000 sites. That would take some tweaking. Once done, various population sizes could be modeled and, unless I am mistaken for the first time, the present results could be reproduced more cleanly.
So there is my offer. Tweak the program. Try it out. See if you can beat a paper out of it, and prepare for a trip to Stockholm. If it is agreeable, I shall send you the $5,000 up front. Do as you like with it. I do not ask for victory, only for an honest try. Be sure to let me know how to get the money to you.
M. Linton Herbert