Thread: What is the name of this type of computing

I'm writing a story about AI but I need some information.

I've been told that regular computers can execute a small amount of commands simultaneously (like around 4), but they are occurring so rapidly that it seems much more is happening at once. I also am aware that what sets conventional computers apart from the brains of animals is that in brains that there can be many sets of neurons firing simultaneously in separate parts of the brain independently of one another. (If I am grossly misinformed about any of this, please don't hesitate to tell me.)

Now, first of all, I'd like to know the name for a type of computing that would behave more like in the brain without such a limited amount of simultaneous processes. Also, I vaguely remember reading an article about some sort of micro fabric that was recently invented that basically could do that, which was being used to simulate complex cellular activity (or something along those lines) If you could find any information about that, I'd appreciate it.

What you are talking about are parallel processors.

Your ordinary computer is not capable of handling more than one task at a time. If you are playing a game while waiting for a download to complete, the computer gives the appearance of doing both at once. It's not. The central processing unit, cpu, is dividing its time between the game and the download. If you are running four different applications, the computer will appear to be doing all four, but you will wonder what's taking so long. Each application must wait its turn.

Similarly, a very complex operation that requires many small tasks to be completed will, again, require cpu time sharing. This is the case with many of today's video games,

Parallel processors, many cpu's, allow the multiple demands of a complex task or of several tasks to be assigned to separate cpu's.

Imagine that you have four sacks of onions to be carried from the packing shed to a customer's pickup truck. Your nephew Bernie can carry only one sack at a time, so he has to make four trips. If Bernie can get three friends to help him, then a single trip-cycle is needed to do the job. Parallel computing is like Bernie and his friends. Each cpu does a part of the job, so the job is completed sooner.

My understanding is that our brains are capable of parallel processing so that different tasks or different parts of a complex task can be handled simultaneously. It's also my understanding that the extent of parallel wiring in our brains is an important element in determining how intelligent we are, or how intelligent we appear to be.

The fabric you are talking about is one of several recent developments in manufacturing technology that allows the mass production of cpu's that occupy very little space. In the near future we will see that kind of technology appear more and more in our home and office computers.

Edit - You mentioned commands, and I talked about applications. The cpu can execute only one command on each clock cycle. But many commands are often needed even for very simple tasks. That is why cpu clock speed is important.

One way of describing brain function compared to computers is "holographic".

Right. The index in your computer has the exact, physical, location of every bit that makes up a file. Even when a file is badly fragmented the cpu is able to issue fetch commands in the right order to those exact locations, string everything together, and deliver the file intact or carry out the instruction set correctly. At least that's what it's supposed to do, and most of the time it all works as it should.

I've read a little about the holographic structure of the brain. It's the opposite of the old idea that the music lessons are stored in precise physical locations, the way a computer stores data. Apparently that is not the case.

I had a massive stroke in 2003 and for a while had no memory of the days just preceding the stroke. At the same time I was able to remember incidents from very early childhood that I had not thought of in years. Gradually my memory of the events surrounding the stroke cleared. Had they been stored in a specific location and that location wiped out by the stroke, then the lost memories would never have been regained.

When computers can be built to withstand damage equivalent to a stroke, then we will be well on the way to true artificial intelligence.

ianfort - Is any of this helping?

One of the best illustrations of "holographic" information storage is this: If you have a hologram, one of those 3-D eyes you used to see, or some such, and cut it in half, you don't see two half-eyes as if it were a piece of paper. You see two images of the complete picture, but each is less detailed.
Thinking about what that means as far as "where is the information" is illustrative. I have a hard time imagining a serious AI that didn't work on hologrpaphic computation.

Thanks guys! This information is all very helpful. Just one more thing: What is that special processing fabric called? I'd like to be able to look it up.

I see it in a quantum sort of way, with millions of packets of information dispersed mist-like through the brain. Each packet of information contains a rough outline of the whole. Each packet has a slightly different set of data, and each, by itself, will give you a picture of the whole, though with few details. Combine two of them and you get a slightly better image. Combine a million of them and begin to get many smaller details. Keep combining and eventually you will have everything the brain has stored for that image.

The picture of the tractor in the rice field is complete enough for you to tell what it is. But look at it closely and you will see there are in fact no fine details. It's a usable image, but not truly complete. I reduced the total number of pixels to make the picture fit the requirements for an avatar.

I'm wondering if our brain does the same thing. I've often wondered how it is that we can keep learning, keep absorbing new information, without eventually running out of space in our 'little attic', as Sherlock Holmes put it. And I believe Conan Coyle put the right explanation into Holmes' mouth. Every time we learn something new, we crowd out something that was already there. We don't lose it completely. It just gets compressed and starts losing detail. Sort of like converting from raw to jpg. Some of the finer details are lost, but the image is more manageable. And then if we keep compressing, as I do with an avatar, we keep the basic image but lose the detail.

For years I developed film in D-76 and paper in Dektol. (No I didn't use any of the exotic formulas. Sometimes it was off-the-shelf Kodak if I couldn't find everything fresh.) But now I'm sitting here realising I do not remember the details of the formula for D-76 or the time and temp chart for Pan-X, Tri-X, and Verichrome Pan. I've been shooting digital for years now, and all the details of film photography have been lost as my memory of that technology is compressed to make room for the instruction set for my Canon EOS.

And do not say what you are thinking.

So an artificial intelligence device would need this kind of capability, to be able to remember the outline without cluttering itself up with the details. And the AI machine would have the same ability I have. I may have forgotten the formula for D-76, but in just a few seconds I found this...
D-76 type Film developers by Ryuji Suzuki
and if there were a need to mix a batch of D-76 I could now do it. Therefore any independently operating AI entity would need access to a vast database to take advantage of its own capabilities.

There is the basis for interesting science fiction. No monsters, nothing illogical. Just an extrapolation of where we are today to where we may be fifty or a hundred years from now.

Reality is a great place to live.

fabric? Do you mean a hologram? Just google it up

No, he's talking about a manufacturing process that allows the mass production of cpu's in clusters on a fabric-like substrate. I'll try to get some information on it.

Make for some fabulous t-shirts

Conan Coyle?