Originally posted by byhiniur
Originally posted by byhiniur
It just raised the issue in my mind of why we would want robots. Fair enough they are cool in movies but apart from that... Its just the quest for
immortality, and it'll only end in infamy.
Not why they want them, but everyone...
Robots will (and are) be(ing) used for war: fact. There are several DARPA and broader DOD programs (AI & machines) and such for just that. As far as
doing labor i think they'll only need them for certain things, dangerous things. It's quite possible that NBIC neural implants could be ready by the
time robots could make themselves useful. The title of the NBIC manifesto is "Converging Technolgies for Improving Human Perforamce:, with a nice big
Dept. of Commerce logo on the cover. The issue will be whethre it'll be more feasible to essentially turn us into the robots, by improving our
performace. That is a goal, along with "collective sociey" hive hierarchy.
I suggest you take a good look around this site:
www.darpa.mil
IIB, your idea seems pretty watertight, I've read your flyers and think some of the issues you have raised are important. I'd raise one
point...
I've just watched the documentry in the matrix box set called the science behind the matrix. It STATES, specifically, that an electric brain would be
much more efficient than a biological one. However big you want to make the brain, if the neurons are man made it'll be much more
efficient.
Of course it would, but you have several implications to consider.
1. Lifespan: "chips" can burn out, or sections can that seriously inhibit the rest of the chip. DARPA actually has a program for having 3d
semiconductors, up 100 even 100 layers. But hat about the power consumption? The heat? What about when a section goes out crippling the chip, the
system goes down. You cant go in and repair silicon.
Nanotech (DARPA Ultra Electronics) will help eventually, but nanocircuitry has a long way to go. It does have a potentially promising future I will
say, but theres no telling when we'll ge there. Self assembling -effective- nanocircuitry is much more difficeult than solid parts. Like NBIC, the
skies the limit in the future tho, but which AI will get us there first?
Quantum is nano in many ways, but the way theyre using it is in its own class compared to standard silicon or the meaning of Ultra Electronics. Real
quantum computing will seriously revolutionize computing, but I expect it to be mostly supercomputing (especially government) for some time. It's
basically teleporting photons around which can have more than just 2 states like binary silicon. It will mean entire new realms in software, and also
in manipulating larger scale brains.
Experts will tell you that for every hour of hardware engineering it creates 24 hours of software engineering, espacially in a task like this. This
begs the question, would it be more feasible to us ethe built in programming of the live neurons?
We still dont know exactly whats going on inside them, and probaly wont for some time, because they are advanced and complicated. Thats one reason why
i dont buy the notion that our brains are only 600MB (Sir Charles) or 50KB (Kurzweil if I'm not mistaken).
Neurons are the primary cells used in our brains for everything from sensory nerves to cognition. They is somewhere between 10-100 billion neurons in
the human brain. They're said to have up to 100,000 synapses. They self organize as they learn, using proteins on the surfaces of the cells. Further
studies suggest that neuron dendrites and axons reverse fire.
Traditional AI schools base their neural networks on the assumption that cortex neurons use action potentials, such as “Digital” electronics,
instead of analog (graded) transmission. For some reason they’ve been holding onto that belief despite it being known that sensory systems use
analog transmission, and even DeMarse's F22 brain monitored both signals.
A science journal report yesterday, titled “Analog Axonal Signaling” demonstrates that neurons use BOTH digital and analog transmission, even
simultaneuously. I wouldn’t be surprised if they can even use more than 2 action potentials as the binary we’re familiar with does. This could
mean entire new types of math computation theories, but I haven’t seen the AI conservatives reactions yet.
The notion that neurons are “simple” math units is obsurd, but that’s been one of the best arguments against the idea of using neurons instead
of nothing but math models.
According to the textbook the “Biophysics of Computation”, by a top neuroscientist named Christof Koch, neurons are more like IC chips than they
are single switches, as most conventional AI neural net people tend to think.
If we had a silicon chip that had as many 'connections' would it become intelligent? No. In brain hardware, the "learning software", "ROM
memory" and RAM are all built in. You don’t need to program those things, you only have perfect teaching the neuron networks and the skies the
limit afterwards. There are certain areas or 'parts' that play important roles in consciousness, that wouldn't exist in a puddle or blob of
neurons, but the fact remains that the power is in those neurons, and we can tap them right now.
Neurons process and they store memories and it’s been shown that individual neurons can assign themselves to individual people. They're not just on
off switches, they store even photographic memories.
It's suggested that glial cells even help electrically 'compute', and it isn't known how significant their function is.
Does anyone think that we will ever have self-repairing silicon chips? Neuron networks self-repair, and self-form, which would take serious overhead
of the software from the hardware.
The human body makes things out of protien... not really a match to copper and titanium."
Actually it'd be more like gold and platinum. As of this week they can make ONE nanometer gold wires. That means far more accurate MEA's, but then
again they can probalby get better results from using 3D or cube shaped silicon walled chamber, they can get great neuron comms with silicon nowadays,
and theyre still more capable than the silicon. There's alot of wats they can go with it these days.
You seem pretty intelligent to have created those flyers and stuff, but as everyone in the other thread said, I think your wrong.
I dont doubt that one day they could more feasibly outdo live neurons, but for now we can tap power that we dont fully understand, but otherwise have
to dedicate serious resources in massive grid arrays to still not match up in even "connections" alone, and it takes much more than connections.
Right now we have some speedy electronics, but our brains prove to out class computers in basically everything important besides extremely accurate
math calculations. Typical human brains barely even compute, it's actually memory associated like language in many cases.
It's all about time. Within 5 years they will have super AI, regardless. It will incorporate silicon, neurons, quantum and some other forms of nano.
It will eventually be tapped into a list of government (and corporate) systems that will make you sick. Take a GOOD look at the world, and this
country, with all of this AI technology + transhumanism and tell me that time isnt grinding down. We're facing beyond scifi AI hooked into mor
ethings than I can name, while decending into WW3 under fascism with class wars like history has ever known. I think it's time people take a good
look at the situation. We cant ignore this many threats, we cant pretend they dont exist and not help ensure that people find out and things change.
IIB
[edit on 9-3-2006 by IgnoranceIsntBlisss]
[edit on 9-3-2006 by IgnoranceIsntBlisss]
