Laws prohibiting corporate stupidity that might result in their being rejected by consumers and winding up bust are not necessary if WE JUST LET THEM FAIL INSTEAD OF GIVING THEM BAILOUTS!!!
I feel better now.
This is flat out wrong. The whole point of a regression is to determine the correlation in noisy data. We're not talking about random points here like paint thrown at a piece of graph paper, but rather a correlation between indep. variables vs. dep. variables which have a distribution. That in no way negates the possibility that the mean values of the samples can be tightly correlated to the indep. vars.
Now what might be the physical basis for high variance in basal metabolism vs. low variance? Well, there are about a zillion parameters in the human body with complex interactions, genetic & epigenetic dependencies, etc. that we barely understand! Yet we assume that everyone is the same?
I'll tell you where this unscientific belief comes from--the "soul" model of human consciousness. Most discussions of obesity have a heavy bias toward the view that people simply choose to be pigs. Evidence that this is false is steadily accumulating, as it is clear that simply turning a few knobs on your hormone regulation, or other parameters, could turn you into a completely different person--an obese compulsive eater, a drug addict, etc. Note that it is easier to perturb a well optimized body so as to degrade health and behavioral regulation vs. bringing one from non-optimum to optimum.
I put my current machine together a year ago with a i4790k, and a decent mobo. Installed LinuxMint 17. Then I did a little test. I ran everything I normally do at once, including VMware with 4GB of potential RAM, running AutoCAD and some other crap, in Linux running LibreOffice, Firefox with dozens of vids playing, mp3 players, Thunderbird, and half of the KDE apps. I think I finally got it near 4GB. So I added another 4GB.
If I was doing heavy media editing, maybe more could help. Or heavy computations, which I don't do much anymore. 4GB would not be an obstacle for general office, browsing, and fooling around. This machine is mostly an indulgence, since there is little perceivable increase in UI performance vs. the dual core 3GHz/4GB one it replaced.
You're asking "what is the system architecture of the brain?"
This would be a good thing to have mapped out. In fact, I'm surprised this isn't done already. Or is it?
Instead of working with amorphous blobs of neurons, the real advances in AI will come when specialized subsystems can be taught, then initialized at will with the "teaching." Those systems put together into a larger system similar to an actual brain will be what leads to general AI.
Byte your tongue.