"No, I don't like it. Turn it Off!"

          -- probably inaccurate Watson quote



I Think, Therefore I'm Silicon

Well, that previous essay turned into a bit of a flame, no? It kind of took an unexpected turn for me, too. I don't really have a problem with the Blue Brain researchers, but I find their smug optimism ridiculous.

Moving on. Can we ever create a conscious entity out of silicon nuts and bolts? First, we have to decide what is consciousness. Is it something squeezed out of our brain's electrochemical processes or is it something we're born with? If you keep an open mind to all the evidence the world provides, both scientific and anecdotal, you could end up making a case for either side.

Serously, there's enough spooky unexplained wierdness going on out there - "I see dead people" - to make you wonder. OTOH, if someone is unfortunate enough to sustain irreversable brain damage, from accident or disease or whatever, their consciousness is likely to go through a change that is also pretty much irreversable. There is some sort of a tight, physical connection between the two. So how do I avoid writing some metaphysical slosh, yet address both sides? You know I have to give it a try!

Order of Magnitude — if you're reading this semi-technical blog, you probably already know what OOM is. If not, here's a quick explanation. If something is 10 times bigger, or smaller, than something else, they differ in size by one order of magnitude. If one's 100 times bigger or smaller, that's two OOMs, because 100 is 10 to the second power. So if the power of 10 = 2, the OOM = 2. The order is the power of 10. Simple. Order of magnitude can apply to any measurable quantity such as weight, volume, length. For what I'll be writing here, I'll stick mostly to length. The tricky part of OOM is that it is just a number. When the order starts getting really large, it becomes hard to visualize the difference being described. For example, the diameter of the visible universe is 26 orders of magnitude longer than you are tall. I know you know that the universe is really big, but I'm telling you, 26 OOMs means that the universe is really, really, really big. Practically unimaginably big:
   If you = 1, the U = 10,000,000,000,000,000,000,000,000.
And that's diameter, not volume. Volume would take the difference up to 77 zeros. Yikes.

Anyway, to paraphrase Mark McGwire, I'm not here to talk about the big. I want to talk about the small, the tiny, the miniscule. But I'll try to get you there in easy steps, to help you visualize more easily. We need a starting point, so here's one: we humans range in height mostly between one and two meters, most of us being closer to two than one. So as a starting yardstick, we can use a meter stick, which is about 10% longer. For visualization purposes either will do, we don't have to be fantical about accuracy. A large beach ball is about a meter high, that's a good starting point too, in 3D no less.

A drinking glass is about one OOM shorter than the beach ball. One of those disappointingly short ones you get in a restaurant when you order a small milk or orange juice. A coffee bean is about one more magnitude down. The thickness of a credit card is roughly another magnitude lower, or three OOMs down from the beach ball. It would take a stack of about 1,000 cards to reach the height of the ball. Yes, that seems high to me too, but take a look at a credit card next time you use one. They're surprisingly thin, especially if you discount the raised numbers.

A human hair is an example of the next step down, one order down from the credit card, 4 orders down from the yardstick, the ball, and (approximately) you. What do we reach at 5 OOMs? Pretty much the limit of what the naked eye can see, and that's with really good contrast. A speck.

Let's speed this up. Shrinking through bacteria, and droplets of fog, and various viruses, getting to 10 orders down we're about at a molecule of water. In other words, if you could shrink yourself down to the size of a speck that you maybe could or couldn't see in the first place, that water molecule in turn would be a speck that you could once again barely see. Water of course is made up of three atoms, the smaller of which is hydrogen, which is another OOM below water. You may remember that atoms are mostly empty space, and to get to the nucleus of the average atom, you'd have to drop another 3 OOMs. However, since a hydrogen nucleus is a single proton, that's yet another order down. We're now at 10 to the minus 15 beach balls.

So, if you did the shrink thing again down to the size of the water molecule, the proton would still be a speck you probably couldn't see (actually, you really couldn't see it because now we're way below the wavelengths of visible light. You'd need X-ray vision at this point. But I'm screwing up the metaphor, so never mind). Shrink again so you're eye to eye with that proton, and its constituent quarks would be the thickness of a credit card, which is also the size of that electron buzzing around a kilometer away. That's right, 2/3 of a mile - the atom really is mostly empty space. And you ... are made of atoms ...

Below this point, size becomes more speculative. Neutrinos, depending on their energy, supposedly range from a thousand to a million times smaller than an electron. And yet, one theory holds that since space has been expanding ever since the big bang, some very old neutrinos must have expanded as well, so they would be almost the size of the universe itself. I dare say we've reached the point where consistent measurement has become doubtful.

Even measuring the 'size' of something like an electron is iffy, since at this quantum level it is as much a wave as a particle. If the electron were a photon, its Compton wavelength would be only one OOM down from a hydrogen atom. Yeah, I don't understand it either. Let's just say that the smallest meaningful (ie, not too 'fuzzy') size measurement is that of the electron or quark. And I'm not too sure about those quarks.

So we have arrived at 18 OOMs down from you, me, and the beach ball. You may notice that this value is not so different from the 26 OOMs of the universe, in the other direction of course. This is to say that the relative sizes from the universe to you, and from you to the electron is only about one and a half speck-shrinks different. And we are not anywhere near the bottom.

Since I recently wrote the magical word "quantum," I'll cut to the chase. According to quantum theory, the smallest length that has any physical meaning is the "Planck Length," with an OOM of minus 35. At this point we've reached what physicist John Wheeler called "the foam of space." In other words, from where you were last (proton), if you then did a mini shrink down to a credit card (electron), you'd still be 3 shrinks away from where you would end up surrounded by coffee bean sized quantum foam!

Remember, each shrink is 5 OOMs, down to a size that is borderline microscopic. So the Planck Length is 4 shrinks down from a proton, or 7 shrinks down from where you are now. If you and some friends were rappelling down to the bottom with a very long rope, and you were going first, you would drop out of their sight seven times before you reached the basement. No matter what kind of visualization games we play, this number is basically unimaginable. We can type it out and pretend we comprehend it:
   0.00000000000000000000000000000000001 meter,
but we really don't.

Put another way, it is practically as far down from the electron to the Planck floor as it is from you and me to the electron, 17 versus 18 orders of magnitude. Translate length to volume and it becomes a one followed by 50 zeros of unexplored space.

Do I have a point? Of course I have a point! I wouldn't make you read through all this crap without having some conclusion to offer (although it's been a heck of a lot of fun to write). My point is: can we be so sure that consciousness doesn't lie somewhere within those 50 unexplored zeros? Even if you want to discount the neutrino levels, there are still Orders of OOMs.

Perhaps there are layers of awareness: sensation, feeling, intuition, creativity, who knows, each Orders apart from the other, yet still fitting in well below what instrumentation can reach today.

Indeed, an unprovable theory. If scientists could investigate these tiny (yet huge) regions, they'd be doing so. Since they can't, these speculations are untestable. I don't feel too bad about that, physics is filling up with untestable hypotheses these days such as string theory and multiverses. At least there's math behind those theories, while alas, there is currently no math of consciousness. That means no easy way to make predictions, let alone test them.

How this applies to silicon based artificial intelligences is this: How powerful will they have to become before they graduate from very clever to truly intelligent? Will they ever? I find it interesting that some particles that physicists have predicted but not yet discovered, like the Higgs boson, have nevertheless been corralled to a degree by experimentation. In many cases, physicists can say something like "it has to be smaller than ..." or "it must weigh more than ...". Perhaps 50 years from now, when electronic brains have an IQ of 100 million (6 OOMs beyond human), Scientists will be able to say, "sentience must be ..." Perhaps by then we will have a mathematics of consciousness.


Androids and Cyborgs and 'Bots, Oh My!

Well before the time artificial intelligences reach a 100 million IQ, they will have started designing their successors and shoot up to a trillion IQ in a matter of months, so they won't need us anyway. "Math? We don' need no stinkin' math!" But let's avert our gaze from this unpalatable future and play around with what physical forms AIs might best exist in (you know, after they've shrunk enough in size and power so they no longer need a direct cable to Hoover Dam).

Personally, I'm sort of stuck on one single form. Disney produced a hokey movie in 1979 called "The Black Hole." I liked nothing about this movie except for two robots named Bob and V.I.N.CENT ("Vital Information Necessary CENTralized" - Lord!). B & V were stacked cylinders about one meter in diameter, floating by (I assume) antigravity. Yes, they were also hokey. Even though the movie came out two years after the first "Star Wars," Disney bypassed the use of CGI and went with puppets - with painted on eyes. Pure camp, unintentional comedy.

Yet the concept has merit. I think a self-propelled robot would best be served by a spherical body, or barring that, a cylinder. This would allow room for many different sensors, and allow them to be aimed omnidirectionally. And since antigravity may not be available for awhile, I would mount it on ... a Segway! I mean, why not? Can't you see this contraption rolling down the street, dodging traffic? It could take over the jobs of bike messengers and probably cause fewer accidents. I'm serious! I hear you laughing, don't pretend you're not. Give it some thought, it's a practical design.

Robots in humanoid form seem to engage our imagination the most, at least since the 1920 Czech play "Rossum's Universal Robots," and probably earlier. There are cyborgs, half human / half machine; generic metal robots; and androids, which look as much as possible like your next door neighbor. As scary as cyborg Borgs and Terminators are, androids could end up posing the biggest problems.

They may also end up being the creepiest. There is a "hypothesis in the field of robotics and 3D computer animation" known as the uncanny valley. It has antecedents in Freudian psychology as well. Basically, it says that if an artifact is close to human, but not "quite right," it tends to be disquieting. The 'valley' part comes from graphing this phenomenon. At one end of the graph is a healthy, friendly human. At the far end there's something innocuous yet still sort of humanoid like a Teddy Bear. You can move from the bear to say, a perfectly sculpted, lifesize statue and feel about the same toward either one. It's just a statue. But move on to a robot that has almost but not quite lifelike voice, or movement, or skin, and we tend to be repulsed, even if we can't say exactly why. The Y-axis of the graph measures revulsion, and it dips at this point. There is real research to support all this, by the way.

I would guess we've all experienced this sometime in our lives, even if we've never visited a robotics lab. For me it's monkeys, or more specifically the great apes. I do give orangutans a pass, but gorillas, chimps, baboons - "Ugh." I know it's unreasonable, and some readers may now hate me, but I've never gotten over it. They creep me out.

Probably too much information on the personal front anyway. Back to androids. If we try hard enough, I suppose we can finally end up with androids that can successfully pass as human. We'll even ultimately get facial expressions right, which is darned hard. But will we be glad we did?

Towards the end of the previous blog, there was a quote about social cognition and empathy. Unless we come up with infant androids that can grow up in a human family, they ain't gonna be human. To fully pass as human, they will have to carefully observe our behavior in many different situations and gradually learn to mimic that behavior, without really understanding the basic 'humanness' of that behavior. In other words, they will be sociopaths. The crown of our AI development may well be the perfect serial killer.



Notes:

I think it might possibly be illegal to upload this essay without a shout out to Isaac Asimov's "Three Laws of Robotics:"
• 1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
• 2. A robot must obey the orders given to it by human beings, except where such orders would conflict with the First Law.
• 3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

In later development of his "Foundation" series he added a "Zeroth Law":
• 0. A robot may not harm humanity, or, by inaction, allow humanity to come to harm.

The ethically minded have since modified and expanded on these. (See for example, Laws of robotics), but nobody has passed Asimov for pithiness.

They are certainly a good idea, but like all laws the question is, can they be enforced? Between trillion IQ AIs and human mad scientist types, they will probably end up circumvented. Still, it might be comforting to future humanity if at least most of the robots they come across are law abiding.

It may therefore be a good idea to try to impose other limitations on robots. I say they definitely should be grounded. No flying robots! If some out of control intelligent beach ball on a Seqway comes after me, I want to be able to escape by climbing a tree. Of course, it will probably be equiped with a particle beam weapon and shoot me down anyway. I'm thankful that I'll be dead before nanotechnology can prolong my life by 500 years so I won't have to live in such parlous times.


The ancient Chinese had a saying, "May you live in interesting times." It was meant as a curse. So far, I've found these interesting times to be a blessing. And the future looks even more interesting. So many things are proceeding exponentially: AI, nanotech, biotech, scientific knowledge in general. Something does probably have to give. I'm not sure a finite Earth has room for all these infinities. The times may eventually get a bit too interesting even for me. But for now, I'm having fun.



Acknowlegement:

I would love to claim that I spent hours and hours of diligent research to come up with all those size examples as we shrunk ever smaller into quantum realms, but most of them came from one source: Scale of the Universe. Frivolity alert: it's addictive!





         Introduction to "The Perfect Language"    —   Table of Contents
         Silicon Brain    —   Previous
         Next    —   Eventually




Started: October 25, 2011