Singularitarian Anonymous

“How you doing, Dixie?”
“I’m dead, Case. Got enough time in on this Hosaka to figure that one.”
“How’s it feel?”
“It doesn’t.”
“Bother you?”
“What bothers me is, nothin’ does.”
“How’s that?”
“Had me this buddy in the Russian camp, Siberia, his thumb was frostbit. Medics  came by and they cut it off. Month later he’s tossin’ all night. Elroy. I said, what’s eatin’ you? Goddam thumb’s itchin’, he says. So I told him, scratch it. McCoy, he says, it’s the other goddam thumb.”

When the construct laughed, it came through as something else, not laughter, but a stab of cold down Case’s spine.

“Do me a favor, boy.”
“What’s that, Dix?”
“This scam of yours, when it’s over, you erase this goddam thing.”

- William Gibson, Neuromancer

It appears that my thinking on immortality catagorizes as “singulatritarian“.  I wasn’t aware of that and nobody bothered to come around and tell me that. I had to read it in Time magazine.  Kurzweils thesis is quite interesting and a heavy influence on my own thoughts on computers reaching near-human intelligence.

Exponential Growth of Computing

The Theory of Kurzweil is simple, yet elegant. It all starts out with Moore’s law, which states that “The number of transistors that can be placed inexpensively on an integrated circuit doubles approximately every two years”. This law has been surprisingly consistent the past three decades or so. Put another way; computational power for a standard desktop computer (or any computer for that matter) doubles every two year. This inevitably leads to a logarithmic graph reaching up, up and further up, towards a computer with near-infinite computational power.

Now take Moores law and plot it on a timeline, adjust a bit for computational powers – calculated in operations per second – and try to figure how many of those operations a human brain does. Here is where the whole thing starts to reek a bit, but I’ll get to that later. Kurzweil assumes that a human brain does about  1000 trillion operations per second, which is just about the speed of the latest supercomputer ( at least according to this article ).

Within a few years ( somewhere near 2015) our desktop computers will reach the computational level of the human brain; but more importantly, as Moores law progresses, we’ll have computers with the computational power of all human brains around 2040’s or thereabouts. This point in time Kurzweil calls the ’singularity’. Any prededictions beyond that point are meaningless because we, humans, will no longer be the most intelligent being on this planet.

“What we do know is that sometime in the early 21st century all of mankind was united in celebration. We marveled at ourselves as we gave birth to AI”
- Morpheus in “The Matrix”

To Kurzweils credit, he has done a lot of predictions that actually came true (among which the prediction that a human would be beaten at chess by a computer around 1998; it happened in 1997).  Although I do believe in simulation of human intelligence, I doubt very much if consciousnesses will come into the machine in 2015 or thereabouts.

Another problem with Kurzweils projection is the hypothesis of “the Quantum Mind” (or Quantum Consciousness) .  Roger Penrose, the famous mathematician, together with Stuart Hameroff developed a thesis now catagorized under the name “the quantum mind“.  It states that the brain does not only function on a molecular level, but also on a submolecular or quantum level. Now I won’t go into that to deep, since this theory has many quirks of its own, but I’ll try and jot down the git of it.

Hameroff, an anesthesist, wondered about people having memories of surgeries they had. Some people could recount details of an operation, while they under narcosis. Maybe, he figured, the mind operates on a lower level then we assume where some quantum effects take place; like quantum entanglement. Maybe the mind keeps together while in a sedated state and can even travel outside the body without losing cohesion. Now I know, this sounds very far out there, but it appears that there is some evidence that the brain actually has some parts ( the microtubule to be excact) that operate on sub-atomic level.

Now a friend that actually scans human brains in an MRI machine for a living estimates that the size of actual quantum processes are of by a factor 100 (at least). That is to say, most processes in the brain take place on a scale that is a hundred times bigger then the quantum universe. Penrose argues however that physicists see the universe either as Newtonian ( classic physics, relativistic physics, ‘big’) or on a quantum scale. Surprisingly little research has gone into places where those two world meet, like – presumably – in the microtubule.

Either way, if the brain can tap into this little known world of extra processing power – 100 to 1000 times smaller – and if the brain is indeed  capable of processing cubits, then we shouldn’t be measuring the human brain against a 1000-trillion-operations-per-second-computer but against something in the scale of 100.000 trillion cubit computer.
Which is to say, a computer that is much bigger then what we will have developed the coming century. I must point out though that Moore’s law is logaritmic in nature, which means we will get there eventually as computing power will increase up to an almost infinite level.

Penrose came to the same conclusion as Hameroff in a much more pragmatic way. Penrose considered Godels theorem, which in essence states that a system can never undestand a system of which it is itself part. So humans can never understand the universe completely as we are part of it. Neither can the brain understand itself, as it is of course part of itself.  Ergo; we can never build a computer that emulates the brain, as we are using our own brains to create it.

Quoting it directly from Wikipedia;

“Penrose presents the argument that human consciousness is non-algorithmic, and thus is not capable of being modeled by a conventional Turing machine-type of digital computer. Penrose hypothesizes that quantum mechanics plays an essential role in the understanding of humanconsciousness. The collapse of the quantum wavefunction is seen as playing an important role in brain function.”

I admit that I had to take a little detour, but I’m getting to my point in a bit.

Remember that I said the human brain will not be running on a computer any time soon, at least not in the coming century because of the quantum processes that might go on in there. That doesn’t mean the brain can’t be simulated on a computer in the next decade or so. That we can’t write a program that pretends to be a person and does it so well, that it fools us all.

As I’ve said here, and perhaps more profoundly in my article about being the first immortals, I do believe that some of us – of our generation – will be among the first immortal beings. Even though Penrose seems to be a terrible spoil sport in this regard.
The question I raise is philosophical; does it matter that we’re not actually simulating a real brain. That we’re merely simulating a person up to the point that is consistent enough with the actual flesh to fool everbody else. ( Or to pass the Turing-test). Here’s how I put it earlier;

“My personality can be uploaded or at least be simulated within 50 years. After I die my children and grandchildren will be able to talk and converse with me long after my death. I will be able to reflect and think about the times ahead and about the state of the world in 2070 and beyond.”

Is the simulation me? Or is the simulation my former self, a mere replica, a construct of a person?

“Now of course my immortal personality is simulated; but the human brain doesn’t make that distinction; if my virtual self has lived for 300 years, the simulation is, arguably, more ‘real’ then my own, fleshy, personality as it has ‘lived’ longer then the original. Simulation is not as bad as most people would think. Baudrillard argues that although we might think different, most of our perception is created upon simulated experiences. And even ‘real’ experiences are nothing more than fragments of real observation added with closure.”

And that is where things get Philosophical; let’s say both men are partially right, and we are able to simulate humans in such a way that is completely indistinguishable to others (to other humans). That would mean that we are probably postponing the ’singularity’, but we are then nearing a Ghost-in-the-Shell-like future (or Asimov-like future, whomever you prefer). We  have to ask ourselves what we believe consciousness really is. A future in which machines we’ll ask us if they do not deserve to be called conscious as we do.

And this will all occur well within our lifetime.














- one = 2