Quote:
| only carried over electronic signals rather than brain signals. |
Which, of course, are the same thing. Both the brain and heart run on biologically generated electrical impulses.
For example, your heart generates a voltage which then goes from the top to the bottom (basically) with current passing through cells on the way down which in turn makes the muscles clench/release. Sudden Death Syndrome (most common heart attack) is when the cells get damaged making the circuit path blocked or slowed. The electrical signal starts to lose it's flow path and ends up going back up against itself and falling into confusion (fibulation) making the muscles twitch all out of order (and therefore not pushing blood smoothly throught the system, but instead back and forth in the heart - again roughly).
So this is one big reason why AI researchers feel this confidence, because we know the human body IS based on ELECTRICAL SIGNALS.
Just like Neural Nets are based on the understanding that in the brain memories seem to be made by electrical signals going between some cells creating stronger bonds (like exercising) while less used bonds leave weak connections. The patterns of these connections seems to be memories at this point.
Take an input (via electrical signals from all the various nerve endings, including ear and eye) and place those patterns against and existing template in the brain in a massively parallel fashion and you possibly/probably get "recognition".
Then on top of that place patterns for other things on the same place and there could be a "match" that is not a true match that "reminds" us of something, thus making us think of something other than what we are seeing, like making an artistic connection. In terms of near match but not actual match, think of a square and circle of similar size. Not literally I mean, but just how patterns can sorta match, yet also be definately different. In this case you recognize what it really is, but at the same time you also are made to think of something else as well.
The big leap in AI work was the recognition that LINEAR thinking would not get it done (this is the "programming" you guys are talking about), and a PARALLEL system needed to be used.
At this point we simulate such a system by doing each individual NEURON linearly for an entire pattern, like scanning a video line rather than flashing a single film frame on the screen. Two different ways of producing a very similar effect (moving visual images).
How computers are traditionally used is NOT an attempt at AI. That is CPUs as a tool. AI takes different approaches to "thinking" that are not necessarily efficient and are therefore not used for current problem solving. It's cheaper to pair a person with a traditional CPU to cover all your thinking basis.
But there are plenty of reasons to think that day is coming, just like landing on the moon. Sci-Fi like AI is asking us "then what" before we get there.