First off, we can't upload our brains to any computer memory so this argument is just philosohical anyway. But, what I don't get is if you uploaded your thoughts to a robot, you still haven't uploaded your "self" with it. So even if you could upload, then after the upload, you would have two consciousnesses going, right? You have the one in your meat head and one in the electronic. So, if you uploaded your thoughts and then killed yourself, then what have you gained? Nothing, you just committed suicide! You would have to be able to upload your true consciousness or your ?soul? or your ?spirit? or whatever. Whatever it is that's totally you, humans don't know what that is yet. If we don't know what it is yet, then how can you be certain it was transferred correctly to another container? So, yeah, you might someday be able to make a duplicate of your thoughts that perhaps even continues on after you die, but it won't really be you. After a number of years such a computer brain would certainly deviate from the original anyway. I mean, if eternal life is being sought after, this ain't it. Further, I think it's a futile effort as duplicating ones self and dieing is not really as gratifying as real and true eternal life. For that matter, instead of duplicating ones self, why not just try to make the best robot consciouness instead of duplicating a faulty human consciousness with all of its faults? But again, since this currently is impossible, this would just be a philosophical debate!

There's an alternative to uploading that sometimes turns up in philosophical debates about the nature of consciousness. What if you could replace just one arbitrary neuron in your brain with a silicon alternative? The silicon neuron performs just like the biological one and all the connections to neighboring neurons are exactly duplicated when it's installed. Would your sense of self remain intact - you're still you presumably. After all, biological neurons die and new ones grow and you're still you, so replacing one with silicon shouldn't matter either.

If you can replace one neuron with silicon, why not a few million more? Or a few billion over time? Eventually, one by one, you replace them all and become 100% silicon. Did you stop being you at some point? Obviously, this is no more feasible right now than uploading your consciousness but it might solve the problem of retaining your identity through the transition. (and you don't have to worry about which you is the real you this way!)

Although I too belive that the time will come in the future when we will understand the operation of the human brain in sufficient detail to eliminate much of the mystery of human behavior, it would be my guess that the functional operation of the first intelligent machines could be as dissimilar to biological neurons as the space shuttle compared with a fruit fly.

Many of our low-level evolutionary heuristics (emotional drives) might not be particularly appropriate (sex, hunger, ...) for an intelligent machine. In addition, most of the human brain is optimized for human muscles, organs, and senses, which could also be of questionable value to a machine.

Although this is largely speculation, I would be most inclined to expect that we might be able to design an intelligent computer well before we have completely decyphered the brain. If that was the case then we might find ourselves in the position where we initially guide the "learning" of these systems, exposing them to information and experiences that would be most benificial.

Once they have constructed a sufficiently sophisticated representation of their world, themselves, and us, we might be tempted to regard them as "functionally" conscious.

Furthermore, these "functionally" conscious machines would have internal representations for their experiences and memories, and would probably be able to distinguish between their abstract concept of the vast, wide world and their limited sensory impressions.

From their observations of us, they might also conclude that we are probably "functionally" conscious, although they might also ponder whether or not we could be aware of the same internal sense of a stream of sensor impressions.

I'm in agreement with you that the first intelligent machines will likely be very different from humans. The transhumanists, I assume, base their arguments on the assumption that machines are being built specifically to emulate human intelligence or to serve as a receptacle for it. I've read of experiments in which they are "uploading" nematodes into a virtual environnment with some success (though the process of disassembling and mapping the nematode's "brain" is destructive to the biological version).

It seems to me that uploading brain data is a huge undertaking not at all a simple process. I mean, I'm not a brain surgeon, but isn't a brain Bio,chemical,electrical,flesy network,etc. We're all wired differently upstairs. To upload the thoughts of a person is more than just scanning for the brain's electrical signals. It's more than just sensing a microscopic 3d arrangement of protoplasms. It's more than mapping out the trillions of neural network connections. It's also an instant in time. I mean, if took you a month to upload the data, it would be corrupted because a brain continues to process and store data while it's alive. If you flash froze the brain then you might be able to capture the moment (yes I'm aware of freezing fish and thawing them to life) but there again the electrical stream would be missing therefore it would probably corrupt the data too. What if you could send in nano-bots to get the data. There again, you'd have to have trillions of them, all getting the electrical, chemical, network data at the same instant. Extracting the data without damaging the contents seems a near impossible task. If you died and then the data was extracted, I expect the data would be corrupted then again too as some of the brain would have died. If you distructively sliced and diced the brain you would certainly distort some of the cells in the process and thereby corrupt the data. Who knows what other hurdles there may be to get ALL of the brain's data. I really doubt we'll have the tech to grab all of a person's brain data in our lifetimes.

My favourite quote from this article:

"There are computers operating on big problems at blazing computational speeds"

Computers operating at computational speeds! Whatever next?

is an oxymoron.. people continue to talk about brain and the consciousness that inhabits it as if the two are somehow separable - which they are not - the ghost is not *in* the machine - it IS the machine.

IF (All there is to human existence is cells) THEN Replacing all cells with their functional equivalent could result in a machine of similar capabilities, with temporal meaning, purpose, direction; and of finite value - like the human original. ELSE Replacing all cells with their functional equivalent could result in a machine of similar capabilities, but with temporal meaning, purpose, direction; and of finite value - unlike the human original. ENDIF

Roger

"Our ability to understand and even reprogram the brain is also accelerating. The latest in vivo scanners can image individual interneuronal connections firing in real time. IBM has begun an ambitious effort to simulate a substantial portion of the cerebral cortex at a detailed level."