The power we need to manage our society and overcome our obstacles is already here in the "white space" between us, not in making each one of us some sort of all-wise genius.
So we need to learn how to manage the white space. In computer science, this would be called a distributed "operating system." The term "social capital" is related to this. "Emergence" and "synergy" and "teamwork" are related terms.
"Non-linear" is a related term. What that term is trying to convey is that the sum of two things is often greater than you would expect, and is much larger than what you think you'd get by adding up each thing separately.
One example that used to be familiar to every human, but is much rarer now, is a property of actual fires made of actual wood burning, or at least real charcoal in a grill. When it gets to that stage where the visible flame is mostly out and the coals are red-hot, it has a "non-linear" property. If you lay out these red-hot coals in a long row separated by a hand-width, they will probably die out. If you heap them into a pile, where each is near many others, they will glow brightly and keep going. You can do this experiment and see this for yourself. This is a "real fact." And it is a very important fact to keep in mind.
Or, if you're more scientific, you can look up the references to "cavity radiators", which say the same thing.
If you take a piece of metal with a hollow space inside it, and drill a small hole into that cavity, and heat the metal until the outside just starts to glow, you'll see that the small hole you drilled is glowing much more brightly than the outside of the metal.Here's a link to cavity radiation, also known as "black body radiation". Here's a relevant quote from that link:
"Blackbody radiation" or "cavity radiation" refers to an object or system which absorbs all radiation incident upon it and re-radiates energy which is characteristic of this radiating system only, not dependent upon the type of radiation which is incident upon it.The brightness is a property of the fact that the hole allows each molecule to "see" many more molecules across it that it would "see" if it were just surrounded by the dozen nearest neighbors it has in a solid form. The brightness arises from this interaction, from this mutual encouragement and stimulation. Well, the math gets way more complicated, but that's the net effect.
It's a property of nature, and it's an important one to be sure you know, and be sure you believe in with unshakable faith. Things together burn more brightly than things separately. The "extra" brightness is a property of the "togetherness", not a property of the things.In fact, the effect of the "togetherness factor" is so powerful that it really doesn't matter what the things are. In the limit, the nature of the "things" simply drops out of the equations, and the "togetherness" takes on a life of its own.
You can read about this. You can try the experiment with charcoal and verify it. This actually works. You can read about today's supercomputers here. This actually works.
So, now here is the huge emotional and intellectual leap: If this "togetherness factor" effect is so real and powerful, why aren't we using it more to solve our social problems?
As I reflected in my last post, why are we still obsessed with trying to make the "things" super powerful (and experts in math and science), when, in the end, what actually matters is the togetherness effect?
In fact, as W. Edwards Deming was so vocal about, our so-called "educational system" seems designed to destroy the togetherness, instead of to encourage it. Our concept of a "school" is a place where students compete to see whose "thing" is bigger, or better, or smarter, and then we can select those with the "best" thingies and boost them up even more with top honors and praise and scholarships and prizes until they have super-hot-shot-great thingies, even better than the Chinese, you betcha. And then, oh boy, then we'll "win" for sure.
Huh? Researchers in computer science gave up that approach 25 years ago, when they discovered that networks were way easier to build than super-cpu's. IBM's "Blue Gene" architecture has 65,000 dual-chip processors all networked together.
Researchers in Artificial Intelligence gave up the idea 25 years ago when they discovered that networks of very simple "rules" produced more "intelligence", way more easily, than some huge "if-then-else" program that turned out to be impossible to write anyway. I was on a panel on Expert Systems in Anaheim at a conference once, at the table with a team that had just redesigned the aiming program for the Hubble Space Telescope. They had reduced one "supercomplex" program with 50,000 lines of FORTRAN to 210 simple rules, which was 1000 times easier to write, and to maintain, and to debug, and took a lot less room to boot. The rules were things like "If the solar panels are out, don't plan to use stars they block as guide stars."
My expert program scheduled students into rooms for the Cornell Business School, while maximizing odds students could get their desired courses, faculty could sleep late, no one had courses Friday afternoons, the women were evenly distributed in different sections of classes, everyone had a section of something with every other person at least once, groups of 20 students had multiple classes together so you only needed to find one friend to get notes for all the classes you missed, etc. It was still being used 8 years after I left, maybe it's been replaced now. I would have never tried to take on those specs with a classical program written in some language like PL/1 or C or even LISP. Using many simple rules instead of one huge program let me write something 100 times more powerful than I could have otherwise.
I picked the idea of using rules because they are so easy to change and evolve over time. If a rule changes in the real world, or someone realizes a new rule or constraint, you just change that one rule in the system and hit the "run again" button. That's it. No logic diagrams or long nights wondering if every possible case has been thought of.
Not surprisingly, in this context, is the fact that biological genes operate exactly the same way. No single gene does one job -- mostly, they all interact with each other and it's the interaction that does the heavy logic and lifting. So, even God / Nature / biology selected this as the design pattern that made sense to use for the long haul. Since I'm a big believer in patterns that transcend any particular level or scale, and can be reused on other levels and scales, I'd say, if it works for genes and it works for computers then it should work for people - or for sure we should look and see if it does or not. It's a really easy solution to really hard problems and we'd be fools not to check it out near the start of our homework assignment.
This is the trick used to pack a human being into a mere 25,000 protein-encoding genes, which is only 5,000 more genes than a roundworm has. That's because the number 25,000 factorial is way larger than 20,000 factorial. It's not the genes that make us human - it's the interaction between the genes. The "program" is in the "white space" between the genes. When our bodies break, much of the time, as with current parallel processing software, the breakdown is "between the lines" not "on a line of code" -- that is, it's the collaboration that's broken.
Note also that our bodies have no such thing as a "master gene" or a "king gene" or a "Rambo gene" that directs all of the other genes and tells them what to do. Fascinating. The same thing is true of our visual system - we have subsystems that detect edges, or color, or texture, or motion, but there is no "king system" that runs all the subsystems -- the subsystems just kind of keep up a continual dialog with each other about "what do you make of THAT?" and the vision "emerges" from that interaction. This is a very, very, very powerful design pattern.
Ditto for large-scale heaps of legacy systems, such as a major medical center might have. All these fragments have to try to talk to each other successfully, and most of the breakdowns are not "in" a system so much as "between" systems. Again, it's the white-space on the chart of "systems" that is the "system" we need to keep an eye on.
I've been pondering this for some time. See Intelligent Agent Infrastructures For Supporting Collaborative Work (Sen, Durfee, and Schuette, 1995 - Computer Science and Engineering graduate project, EECS Department, University of Michigan)
In radio astronomy we can find other examples. The huge 1000 foot "dish" antenna at Arecibo, Puerto Rico, was great, and I have a spot in my heart for it since my then room-mate at Cornell designed the central antenna, but times more on and arrays of many smaller dishes are more powerful and more flexible and way cheaper to build. It's all the same principle. Many things, working together, can always trump one big thing. Always. Supercomputers. Intelligence. Perception of the universe. Always.The same thing is true of sports teams, or of any kind of team. The teamwork factor, or what I'm calling the "togetherness effect" can be more powerful than any single superstar. Great coaches look for good players that, above all, are good team-players, not egotistical superstars.
The "Rambo" model is dead, or should be. It fights back and tries to stay alive, tries to keep its grip on our consciousness, tries to keep us mesmerized in believing in it.I could speculate on the role of male psychology and biology on this historical obsession with whose thing is larger, but the honest to God truth is that size doesn't matter. What matters is the togetherness effect. If we could advance past the 1600's into the 21st century and recognize that, we could solve our social problems instead of being perplexed and baffled as to why even our best schools aren't putting out students with thingies so good that they are solving our problems for us.
I mean, come on, guys. Get over it.
"Group work" should not be some sort of after-thought that is tacked on to our "education" if there is time at the end. Group work, and how to make groups work, should be the focus of our entire educational process. Everything else can be "tacked on" if there is time at the end.
There are physical laws involved here. There is no way to make a single computer chip that is so powerful that you can't do ten times better with a network of much smaller chips. Or 100 times. Or an unlimited factor times. The "up" part is in the interaction, not in the thingie.
We have our educational system entirely backwards. That's what Deming said. I agree.
Pouring our national treasure into trying to generate a few more students with super-fantastic math and science understanding is not the answer. "These are not the droids we're looking for." If we did succeed, say, which could take 30-40 years, and every American could score a perfect score on the SAT exam, it wouldn't solve our social problems. Look at Oxford, as I mentioned yesterday - a community of brilliant scholars and they can barely manage to keep the place operating, if that. We'd succeed in making a bigger one of those. Big deal. Whoopie.
Look, reflect, and learn. This is a curve. The road turns here. The future is not in the same direction we were going, even though it is the same road. The cheese has moved.
I just don't know how to say it more vividly. If you open your eyes it's right there. The answer is right there. We can pull our heads out of the sand of cultural depression about the size of our thingies, and get on with life, and solve our problems, and go explore the stars and find really neat stuff.
Also known as "God has a wonderful plan for your life." Look at the center of the neighboring galaxy, where the interactions are millions of times more powerful than "they should be", and feel the power that is right here, waiting for us to accept. There is hope, after all, for even us.
Photo credits:
M31 (Andromeda) Galaxy, Arecibo and Very Large Array radio telescopes are all from Wikipedia.
Cavity radiator pictures are from the textbook Physics by Halliday & Resnick, 2nd ed.
No comments:
Post a Comment