Kurzweil’s Singularity Envisioned by Masamune Shirow

A couple years ago I read read the Stand Alone Complex novels by Junichi Fujisaku and blogged about it on Nov 11, 2010. I wanted to re-post the piece here with some updates to the content, so this is a completely new revision. The novels were like episodes of the Ghost In The Shell: Stand Alone Complex animated series, which is fitting since Fujisaku was one of the lead writers of the show. This is derivative from the original comic/graphic novel by Masamune Shirow.

There are some minor things lost in translation but you get the gist of the story. The reason why I got interested in GITS in the first place is for Kurzweilian singularity philosophy. GITS is perhaps taking place AT the singularity and soon after, but if you extrapolate technology ahead 15-20 years, it may be seen as evolutionary advance, progress in technology with the mapping of the brain and full understanding of neural signals so that they can be intercepted and returned artificially.

For anyone not familiar with the original Ghost in the Shell movie or the Stand Alone Complex animated series (which ran for two seasons), I recommend starting with the post-2nd season movie, Solid State Society. In my opinion, if you start with the original movie you will be confused by the SAC animated series which differs in many ways. Solid State Society is the third movie, and based closely on the animated series rather than the two prior films.

Solid State Society is a fun romp through a technological near-future without all the creative imagery and confusion of GITS, which among other things, is confusing and not intended for children. It is multi-layered, not to be taken at face value, and that’s the artistic aspect–like dropping a pun on someone who doesn’t get it, and not explaining it to them for your own amusement. I sense that kind of directing throughout.

Solid State Society is not quite as complex, with only two layers that are explained. You will benefit from seeing the TV series first, because there are some spoilers otherwise. Anyway, this isn’t meant to be a review, but an extrapolation… What I see in SAC that makes me consider it post-singularity is the mapping of not just brain patterns, but of the entire consciousness. It is magic.

(See the related news article about GHOST IN THE SHELL: ARISE!)

Arthur C. Clark famously said, “Any technology, sufficiently advanced, will seem like magic.”

We can’t imagine transferring consciousness into a computer because our computers are primitive batches of processors. The future computer that will house consciousness is like this… imagine the WHOLE INTERNET today… circa 2013… all those billions of web sites and everything.

Compress that down into a single computer system. How large and complex would that computer be? Now compress it down into a single CHIP.  There you have the technology of SAC, circa 2025. That’s only 12 years from now. (The later movies extend the date into the 2030s).

The whole internet on a single chip… or say, little circuit board, little motherboard, like a flash drive. How much processing power and memory is that? Kurzweil does the math in his Singularity book but let’s just say it’s the power of your iPod or PC times 1 billion. Is that fair? times 10 billion? And on a single chip with processing and memory combined. So what computer power are we talking about?

An iPod/iPhone/iPad is about the same performance as your average consumer PC. X2 for a power user. X4 for a developer or gamer. X8 for a serious hacker or scientist. So we’re talking about 1 Gflop. Multiply by 1 billion and you get 1,000,000,000 GFlops. See www.top500.org. But, that’s also rows of rack-mounted systems.

The #1 supercomputer in the world at the time of this original writing (11/11/2010) was a Cray XT5-HE, with 224,000 cores. That’s 37,333 6-core AMD Opteron CPUs @ 2.6GHz, each of which does 10 GFlops. These are rare systems in the computer technosphere, with only 500 measurably worthy supercomputers out of about a billion computers in the world.The Cray does 2,331,000 Giga (billion) flop/s at peak, which is 2,331 Tera (trillion) flop/s, or 2.3 Peta (quadrillion) flop/s. It’s highly specialized; you couldn’t do your taxes or play a game on it, but you could simulate all of the air particles in a hurricane. That’s equivalent to about 2 million regular consumer PCs.

Today, as of the latest Top500 list (11/2012), the #1 is Titan, a Cray XK7 system installed at Oak Ridge, which achieved 17.59 Petaflop/s. Titan has 560,640 processors (AMD 16-core Opteron chips), including 261,632 Nvidia K20x accelerator cores. So, in just two years, the fastest supercomputer in the world increased performance from 2.3 to 17.6 Petaflop/s–that’s an astonishing 765% increase in processing power. 7X increase in 2 years is 3.5X per year, which is right in line (if not exceeding) Moore’s Law. However, Titan is a huge computer that consumes 8 megawatts of power, and what we want to stick to is a $1000 comparison on computing power (following Kurzweil’s example). Nevertheless, it’s interesting to note the uber-upper-high-end.

If we look speculatively at an “Internet on a chip” in 15 years, it might do 1 billion GFlops–that’s a billion billion flops. Scientists prefer magnitude values, so it is expressed as 10^18 flops. Does this line up with Kurzweilian prediction? Let’s check:

Moore’s Law based on a $1,000 PC. Current average processor, the 6-core Opteron, at 10 GFlops for about $300. Bump it up to $1000, and you get 30 GFlops. Yesterday’s supercomputer for a g-bill, not bad. I’ll round up now and then to make the numbers simpler.

60 GFlops for $1000 by mid-2011
120 GFlops
250 GFlops
500 GFlops
1,000 GFlops by 2015

So, in five years, we’ll reach the first Teraflop consumer device for $1000. This might be a next-gen smartphone or iPod music player type device, or a typical business PC.

32,000 GFlops by 2020
1,000,000 GFlops by 2025
32,000,000 GFlops by 2030

We’re shy of the 1 billion GFlops by 10^3, which is only about five years off the mark, and it’s very likely that technology will exceed these estimates. GITS and SAC were created by an artist, not a techno-philosopher, but he was pretty accurate.  I think he came darned close to Kurzweil’s singularity. But, how does 32 million GFlops equate to the singularity? It doesn’t alone, but it’s the processing power we can estimate in a $1000 PC by 2030 that matters. What about the developer, gamer, or hacker systems, not to mention servers? I’m just talking about an iPad type device at this price level.

What makes SAC intriguing for anyone who speculates about the Kurzweilian singularity is the world described by Masamune Shirow in an early 1990s graphic novel, a whole decade before WiFi became part of the social consciousness, let alone affordable and widespread. Who would have thought in 1995 that WiFi would be free in many cities? People talking to each other through the air back at the time GITS was published was mis-interpreted as telepathy at the time. What is telepathy? MAGIC.

The ability to talk to others with your brain waves. Is it really telepathy? Well, YES, actually, it is exactly that, communicating thoughts to others without speaking vocally. The thoughts are transmitted to others in a shared comm port like a chat room, but at the time GITS was published, most people were not on the Internet yet, and those who were used dial-up modems. The films Sneakers and Hackers came out at that time.. watch them for insight into the time period.

How did Masamune know that we would be using WiFi in the future to communicate mentally, let alone all of his other highly speculative inventions like cyberbrains? He’s a creative genius, plain and simple. Kurzweil might be a fan for all I know. Uh, oh, I may have stumbled upon your inspiration, Ray…

At any rate, I watch SAC and enjoyed the SSS film very much, because it is a plausible future. Scary, in many ways, and like magic to our understanding of the brain and technology today. We imagine scenarios like Skynet and the Cylons as possible outcomes. Kurzweil believes we will upgrade ourselves and A.I. will not destroy us, which coincides with Masamune’s fictional future Earth. He also speculates that the American Empire and China will be at war or will be the two great superpowers, which also seems close to the mark in my view.

What will it feel like to leave your body and enter cyberspace through a network port, without losing consciousness? Will your soul go with you, or will just your senses be exploring that outer world? According to Masamune and Gibson (who deserves some credit here), you can dive into a network or system and be killed but your soul never truly leaves your body and enters that system.

However, many of Masamune’s characters have already left their body and exist in a cyberbody, so diving a system is of lesser import. When motor control  and body maintenance functions are removed from a human brain, replaced with machinery such as an artificial heart, lungs, kidneys, liver, what is left but consciousness, critical thinking, and memory? What, indeed–that is the point. A cyberbrain can intercept the signals coming from organs, and send artifical signals to the organs, and they will continue to function.

Motor control is next, and breakthroughs have already been made in the field of medical prosthetics so that thoughts can control artificial limbs rather than muscle sensors–that technology is already in the works for full neural connectivity. Where do “WE” exist, then, in a brain when all of those things have been replaced with an artificial master control center–a cyberbrain?

We obviously think, make critical decisions, respond to bodily needs with action, but are not consciously maintaining our bodies–we don’t have to think about breathing. Motor control is a similar sub-region of the brain. We really don’t need to think about walking,

we just go, although we can make a more conscious effort to walk a certain way or stop if we want, by paying attention to our legs, but that isn’t necessary for walking.

If you note a spider or mosquito on your arm, you don’t think twice about swatting it away, it’s a subconscious response, instinctive. Your consciousness CAN interfere if you have trained yourself to override the instinctive responses, but only if you wish to. So, some brain functions are autonomous or auto-reactive, while some are semi-autonomous, and higher thinking is fully conscious.

What must it be like to be at least partially cyberized? How would it feel if you sent a signal to lift your left index finger, and it caused a little hobby robot sitting on your desk to lift its hand instead, with nothing but your thoughts? Now imagine that gizmo attached to your body? Given enough improvements, year after year, and new innovations in technology continuing to improve, what might NOT be possible by the year 2030?

Will you still be YOU if your wet brain is replaced with an electronic one, and all of your memories and thoughts copied into the cyberbrain? I personally believe that one cannot be copied and retain consciousness–that’s like teleportation/transporter theory, which I feel is fully destructive. But what if you could move your awareness fully into a cyberbrain in order to control your body, like entering a space suit or putting on clothes. Might not your mind actually move into that new area, leaving behind the old?

In Masamune’s future, a 100% fully cyberized person still has full rights as a citizen, even when they are 0% biological. The heroine of GITS and SAC, Major Motoko Kusanagi, is such a person, and she has the ability to not only change her cyberbody for another one, but she can control two at once over the net (something Masamune says no other human can do). What must it feel like to have no subconscious or motor control aspect of your brain left, only higher thinking ability? Would it feel as if part of you is missing?

That will be the invention of A.I. Not by artificial means, but the transference of a human consiousness into an artificial substrate. Interestingly, that is how the Cylons originated in Battlestar Galactica.

Leave a Reply