We can quickly see that our intelligence is rooted in our physicality. The biggest problem I see with most modern AI work is that it relies on a non-coporeal intelligence.
If you give a robot very minimal intelligence, but give it a body it can navigate around objects much more quickly than if you give a robot incredible intelligence and have it map out a way across a room on a theoretical level and then move a dummy bot across the real terrain.
Out of Control by Kevin Kelly is a pretty good lay resource on the topic. I disagree with him on just what emergent behaviours are (I think that we can get everything from first principles, he does not – but we’ve accomplished that with water, so that is a nice feather in my side’s hat).
Intellect is rooted in the body. Remove the body and what you end up with is a calculator. Human intellect is an outgrowth of the intellect seen in other animals. Our intellect is rooted in our physicality.
Think about it, I could explain the way my apartment looks to you in such a way that you would have a pretty good intellectual understanding of it. If I then gave you a blind robot to navigate around the apartment, in a race a blindfolded me will win 100% of the time – and that is assuming you have a map and I don’t! Heck, a blindfolded random person who has never been to my apartment would win given much less information than you had. Alternatively, I could explain to you certain techniques that I perform at my job. While you would have a very solid intellectual understanding of what they are, you would have great difficulty doing them the first few times simply because the real knowledge of how to perform them is rooted in the physical – indeed, many things really only start to make sense once you do them because the intellectual explanation of them is far too difficult. I could explain how riding a bike works based on the theory of interia, and momentum transfer (as well as how gears and all that work), but none of that will actually help someone learn how to ride a bike.
Indeed, We’ve already created non-organic creatures with intelligence on the level of insects (they say ‘ants’, but the robots aren’t social, so beetles might be a more appropriate description). No reason why we couldn’t take that further. This isn’t a case of intelligence from non-intelligence, but rather physicality manifesting itself in action. Intellect is action in a system with many degrees of freedom.
Given that we already have matrices that ‘dream’, I think it is more than fair to say that these indirect thoughts are just a means of clearing out the junk we carry around in our mind. It is akin to defragmenting a hard-drive. It has already been documented that if a person is prevented from dreaming they go pretty crazy pretty quickly. That makes good sense, since they are trying to hold everything in their short term memory and there just isn’t enough room for it. Continuing in the vein of materialism, I would propose that we treat thoughts as ‘things’, because that is what they are – manifestations of controlled processes in our bodies. Think about it in terms of signaling cascades – there is a signal, a pathway that the signal travels, the means whereby the response can be actualized, and the response itself. I think of thoughts as the means whereby the response can be actualized. A step in the process, if you will.
The ‘mind’ is nothing more than what we call the network of the brain, in the same way that water in a glass is ‘water’ and not just ‘H2O’. So, we could get that world-spanning conscious going if we could, a) properly link everybody’s brain together, rather than having them as separately operating binary units and b) could given that brain some sort of a body to interact with.