However, there is some merit to the claim that at least some of the criticism of the idea of artificial intelligence is almost religious in nature. John Searle is one example of a person who seems to be certain that the phenomenon of consciousness is dependent upon its physical realization, meaning that any implementation of the same processes in another medium is simply a simluation, which could not be considered the real thing. Exactly wherein the difference lies between a real instance of a phenomenon and a mere simulation of it, is something which I have never seen an adequate description of from any of the skeptics. It would seem though, as if any such description would have to include either the claim that there are aspects of the phenomena of consciousness which transcend the physical processes, or that the essence of any physical process lies in something else than its structural properties, which seems to me to be indistinguishable from mysticism.

After getting a hold of the technical documentation and reviewing the requirements, I have to say that you and George are absolutely right. IBM did not in fact model a feline cortex despite their claim. A full post is in the works to explain exactly where the Blue Matter team fell short and how.

In addition, when I was talking about transhumanists, I wasn’t referring necessarily to you or the entirety of the Singularitarian movement since I know full well that not all Singularitarians are also transhumanists. My link to your comment about the functions of the brain was just to illustrate the frequency of this idea, nothing more.

It is quite false that transhumanists are up in a buzz over this. In fact, leading transhumanists like myself and George Dvorsky immediately pointed this out as hype:

http://www.acceleratingfuture.com/michael/blog/2009/11/jamais-cascio-i-can-has-singularity-the-need-for-precision-in-ai-reporting/

http://www.sentientdevelopments.com/2009/11/ibms-claim-to-have-simulated-cats-brain.html

It’s unfair that leading transhumanists go to all the trouble of being right about this and get no credit from you.

As for the idea of creating AI by simulating a brain, it’s ultimately an empirical question. There is no way you can put 100% probability on the hypothesis that a sufficiently detailed simulation of a human brain does not produce intelligence, which is what it seems like you’re doing here. The Whole Brain Emulation roadmap published by the Future of Humanity Institute at Oxford addresses the need for chemical details and describes detailed methods for determining the precise chemical makeup of the brain. You call the project a “decadent” one, and you’re right, but without knowledge of the precise scanning tools available and how long they take to scan, you’re engaging in a sort of hand-waving.

The details aren’t really dismissed in this contention. If you read the Whole Brain Emulation roadmap, you’ll see that it is over 100 pages long and has hundreds of detailed references.

http://www.popsci.com/technology/article/2009-11/blue-brain-scientist-denounces-ibms-claim-cat-brain-simulation-shameful-and-unethical

However, a quantum computer would be used to solve unusual problems and to get a major speed boost over the machines we have today and the result will still be in ones and zeros. I’m much more bullish about future microprocessors being made with diamonds to overcome the current limitations of silicon chips.

Still, both approaches are about doing what we already do, only faster so we can take on bigger projects and make more complex models. The problems discussed in the post will still be there unless there’s a biological element to computers. Until then, the best we’re going to get a stream of data which will be very useful, but it won’t be AI by any means other than a stretch of the imagination.

Now hooking up a real cortex to a computer is something I’d like to see and it would be a huge leap in cyborg technology. We know we can do it because we’ve hooked up computer chips to working brains already with very promising results.

Is it possible and what would the implications be concerning an AI?