Python Liberation Front

Artificial Intelligence has a long, long way to go to catch up with Natural Intelligence...

I have read several books like Kurzweil's "The Age of Spiritual Machines" that put forth the notion that computers will achieve human-level intelligence in 2-5 decades or so. Having thought about this a lot, I disagree, and so I have put my thoughts into the following little essay...

Artificial intelligence scientists are a long way from matching human intelligence. Some speculate that when the number of transistors in a microprocessor equals the number of neurons in the human brain, rough equivalence will prevail. But this ignores the fact that the human brain is three dimensional, with numerous connections to and from each neuron, and with both electrical and chemical processes occurring over each connection; and more importantly still, these electrical and chemical processes are both analogue, continuous variables that have an infinite number of possible states and values, compared to the simple on-off binary state possibilities of digital bits.

Thus, even when the number of transistors on a chip equals the number of neurons in a human brain, the chip will be many orders of magnitude more simple than a human brain.

On top of all this, the brain has many diverse subsystems, and is integrally connected to a whole body nervous system, and indeed to the flesh and blood of the whole bodily system, including its sensory and motor organs.

Of course, what of the hardware-software comparisons between computers and human brains? It might seem easy to equate the brain to hardware, and the mind, or thoughts, to software; but this is a very inexact metaphor. The brain is like hardware alright (although many orders of magnitude more sophisticated hardware than the most advanced computers); but the thoughts and mind of a human are only very superficially like software. Computer programs are determinate, dumb brutes that execute logical commands in a sequence or that interacts instinctually to human inputs. The mind is an open ended system with its own free will, emotions, drives and creativity.

It will take quite a lot to duplicate human level intelligence. On the hardware side, it will be necessary to come up with entirely new levels of complexity, with numerous intelligent subsystems all controlled and orchestrated by a logical mind with equivalents to emotions, drives, willpower and randomness. On the software side, though, is where the real work lies.

Software will need to be developed that is object oriented in the extreme, with individual objects having their own innate goals, drives, priorities, emotions and that are orchestrated in a "society of minds".

This way lies progress.

"The Probability of God"

"The Probability of God" is a new book by quantum physicist Stephen Unwin that attempts to calculate the probability that God exists by using Bayes Theorem. Starting with a 50-50 chance, he proceeds to use six cirteria that he evaluates subjectively as evidence for or against the existence of a God such as the one believed in by the Abrahamic religions.

If you like reading books written by a quantum physicist who cracks a joke every other paragraph, then it is worth buying this book. Otherwise, you can read it at the bookstore as it is a relatively short 200 pages and it scans very quickly.

You will learn the basics of Bayesian analysis, which is interesting. If you are also interested in philosophy and religion, so much the better.

For what its worth, Unwin calculates a 67% probability that God exists, leaving a 33% chance that He does not exist, a gap that must be filled by Faith.