Friday, March 4, 2011

Ray Kurzweil's Singularity Theory & Quantum Consciousness




I wanted to comment on Time Magazine's recent article titled "2045: The Year Man Becomes Immortal" but it seems they are not accepting new comments anymore. I, hence, am using my blog to state my reactions.

First of all, let me say that I'm, in no way, opposed to Kurzweil's theory. As a matter of fact, as a technology enthusiast, his idea that machines will, one day, evolve to become conscious, does not seem far-fetched to me. My generation has witnessed so many advances and achievements in the various fields of science and particularly IT in the last 20 years that it would seem rather preposterous to say that this unstoppable march might just stop when it almost reaches human intelligence. Such a narrow view would, indeed, not be smarter than when our ancestors thought the Earth was at the center of the universe.

Unfortunately, I don't think I can agree on the numbers that Ray Kurzweil takes into account. It seems to me that his whole theory is based on the premise that consciousness or human intelligence or whatever you wanna call it is simply the result of an astonishing amount of neuronal connections. Based on the amount of connections in a regular human brain, Ray Kurzweil estimated that machines will have enough connections to become conscious by 2045.

An easy way to criticize his theory would be to point out the petition of principle in his reasoning. He is assuming he knows what consciousness is all about when he demonstrates how machines will achieve it. The only problem is that we are yet to find out how consciousness arises, what it is and how it works.

Another critical approach would be to ask the following question: what if human intelligence weren't the simple result of an astonishing amount of neuronal connections? I'm not necessarily talking philosophy or religion here. I might just, once again, mention Stuart Hameroff and his quantum consciousness theory. To him, consciousness isn't the result of a huge amount of neuronal connections but would be based on the principle of quantum computation in microtubules within neurons and this dramatically changes the order of magnitude to reach consciousness through network connections. Here are the numbers:

- The neuronal connection approach breaks down like this: we have 10^11 brain neurons with 1000 synapses per neuron switching in the millisecond range of 10^3 operations per second. This amounts to 10^17 bit states per second for a human brain.

- Stuart Hameroff's approach takes into account that each cell contains approximately 10^7 tubulins, then nanosecond switching in microtubules automata predicts 10^16 operations per second PER NEURON (you have to take into account we're talking subunits of microtubules being able to change states). Since the human brain contains 10^11 neurons, nanosecond microtubule automata amounts to 10^27 brain operations per second!

So, in my book, it's gonna take a lot more time to reach the so-called Singularity than in Ray Kurzweil's calculations!

Burgerless food for thought!

No comments: