Thoughts and Theories From Roger Andre
for the curious, and people getting to grips with computing and computers.
Saturday 22 November 2008, 6:13 PM
IBM To Create The Great Brain Thinking Machine.
One of the goals is to chase the minds ability to integrate cross sensory information, as well as time/space/object data. The process undertaken here is to reverse engineer the dynamics and function of the human brain.
Much has already been learned about how to create simple wiring diagrams for "brains". The technology exists now to create structures that match the density of neural and synaptic networks as they are in real brains. Not to forget that we have already been playing with neural networks for some time, thus creating an AI that has the ability to solve problems and act as though it is learning.
The brain is more of a synaptic network than it is a neural network, adapting the neurons as it sees fit for any given function.
I wonder if we'll be asked to spare some of our own computing capacity to help out on this one. These things get can get very expensive in the long run.
At some point things my start to move so quickly that it will appear as if everything happened yesterday or even a week ago. The approach may seem fairly slow now but when the machines get involved in the process and this will happen once everything the researchers figure out gets put into the great computational database asigned to the job, then things will happen at lightning speed. I also figure that some resposibility will be given to machines assigned to these type of tasks to order materials they need by themselves and even tell the truckers where to get the cheapest fuel, thus saving money that could be ploughed back into the project.
All this computing power may come in very handy when it comes to keeping tabs on the growing cloud computing networks out there. Your future laptop may even come with a mouth on the side so just be careful not to feed it too much rubbish. The odd salad will work wonders for its internal health.
PreviousComments on this post
Practically all of the machines taking over the world in Sci-fi stories have them either becoming meglomaniacs (Forbin Project and the Cylons in BG), psychotic (Hal 9000) or genocidal mass murderers (Terminator series). My guess is that the first self-aware AI system will be more like Mr McKinnon. It will be self-absorbed, obsessed about something and a major pain in the butt. Keep a good grip on the power cord, you'll want to be able to pull the plug on it!
Ha Ha the last thing we want is for them to become too human now eh?
Big Blue has an interesting "Autonomic Computing" resource at http://www.research.ibm.com/autonomic/
I've met the guys who talk about this stuff and it's interesting (I hope) to learn that they way they talk about self-healing systems kind of mirrors the way the human body has an autonomic nervous system.
I think the theory is that some ailments (in the body) will be repaired by white blood cells - and some IT problems can therefore be self repairing... but some problems necessitate a visit to the doctors or hospitals - this is when IT calls in technical support.
AdrianB
Hey that's cool, I'll have to check that tommorow at lunch with my new acer 1111111111111111111111111111111111 Oh msut be tired, why didn't I know about that damn radio button! goodnight all.
