[TriEmbed] Self Aware Computing

John Vaughters jvaughters04 at yahoo.com
Tue Dec 17 09:19:58 CST 2013


Based on our short discussion yesterday concerning google and all the talk in the news about skynet and computers being self-aware, I would like to discuss this topic.

We are VERY far away from making computers self aware based on one simple fact:

We currently do not even understand what self aware means in the Human brain.

Given this minor inconvenience I would like to point everyone to a very interesting project called Blue Brain. This project can give us some focus on the path to self aware computing as the leading research into this area. I recommend you look this up if you are interested, but I will summarize to make some points. This project is simulating a brain at the molecular level. This is a fascinating idea because it is within our reach. We can simulate many things in nature by using their fundamental physical  properties. We are quite successful in  many areas with this type of simulation. So this idea that they can simulate a brain by simulating nature in a computer is the real deal, but as with any people that know history of technology, we more often fail in our early attempts to understand complex subjects.  

In 2007 Blue Brain was successful in simulating a strand of rat brain and the results were fairly accurate to a real brain. They are attempting to create an entire rat brain by 2014 and an entire human brain by 2023. I have no idea where they stand on reaching those goals, but that alone can give you an idea of a time frame on self aware computing. In short we are VERY far away if you consider that they reach their goal, the BIG question is........

Will a simulated brain be self aware?

Back to the problem at hand: "We currently do not even understand what self aware means in the Human brain."

Soooooo the big question is will this simulated Brain tell us the secrets of life like as in 42 (Thanks for all the fish). Highly Doubtful!!!! Even if we successfully create this brain, it will be like a babies brain or even worse. A BLANK Brain. It will have learned nothing and will need to be taught something. How do we even teach it becomes the next question and from here on I can think of many problems that we could spend discussing for hours and hours if not days and even more.

What I find absolutely amazing is the amount of computing power required to create this simulated Brain. We are talking HUGE supercomputers and in case you were wondering why it is called Blue Brain, IBM is the computer supplier. I do not have numbers for you, but when I read them in the past I was astounded. It made me appreciate the computing power in our head. It also made me think of something much more frightening and far more possible than self aware computers and that is biological computing. Why reinvent the wheel. Just culture brains and put them to work. Sound crazy??? Well it is, but still more realistic than a self aware computer. By the way, if you are not familiar with the idea of using brain cells as computers, read up on it. They have taught a collection of rat brain cells to fly a simulator successfully without crashing and that was back around 2005ish?

Have I made my point yet??? Technology takes time and just like Dolly the sheep, we fail to truly understand nature in gross miscalculations that take many decades to understand. Rarely does a grand experiment result in a eureka moment.

Having said that, I will now say what I say to all people who fear the self aware computers. Having watched a many automated machines flip out, and corresponding technicians follow, the proper response to any device that is behaving badly is to kill the power.

Anything we create, we can kill. 

That may not always be true, but it is a very difficult task to not be true.

I have explained this to many non-technical people in the past so I am curious to any response to a technical crowd. I am sure I am singing to the choir here.

Thanks,

John Vaughters
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.triembed.org/pipermail/triembed_triembed.org/attachments/20131217/913c7f89/attachment.htm>


More information about the TriEmbed mailing list