Kevin's Watch

In 1965 or thereabouts, Gorden Moore of Intel propounded the above 'law' which stated that computers would double in power and halve in price every 18 months. It may not hold up exactly, but an example of how close he got it would be the following. As recently as the 1990's experimental data at CERN was analysed by the Cray X-MP/48 supercomputer purchased at a cost of $15 million. That machine had half the computing power of a modern x-box 360 available from retailers now for a modest $200.

Clearly this cannot go on ad-infinitum (Moore himself has said as much) but the questions I ask,not adressed in the article from which the above is drawn are what are the consequenses of this exponential growth failing - the article's author implied some pretty serious ones but did not illuminate - and what are the consequences of it not failing. ie What can be done, what changes to our lives can we expect to see, as a result of such unlimited computational power.

I'm kinda a one trick pony when this and related things come up, cuz I always go first to: immortal cyborg human race.

I assume there is a limit somewhere to the process [oh, and left out is the shrinking in size] but not any time soon...and I'm not sure it would matter. Because I think for a lot of reasons that those limits on computational power
capacity exceed the human capacity for using them except in the most extreme endeavors. I mean...how many people really use even 1/10th of the computational power they have on hand? And of that 1/10th they do use, how much of it really serves any need other than convenience or fun?

Well, strictly speaking, Moore's law is that the number of transistors which can be placed on an integrated circuit (inexpensively) would double every 2 years, and he further predicted, (in 1965) that it would continue for at least ten years.

It has, and further than that. However, it's apparently expected that it won't continue past 2015, or 2020.

If I recall, it's not that there's a problem with the law as such, but that we're limited by the physical properties of the materials, and by that time, we will reach a point when the stuff we make the chips out of (currently) just won't be able to handle any more.

Currently unforeseeable advances in the technology may (probably will) change that.

The implications? Computers won't get much faster I guess. Not at the size they are now. Maybe we'll go back to big house-sized ones.

--A

The article I read was by an author called Robert Harris (of 'Fatherland' fame) who, while researching matereal for his new book on the financial wizards at the top end of Wall st, discovered that maths and physics graduates are being employed (rather than traditional economists etc) to produce ever more and more predictive tools that can forcast human behavior to the nth degree. He seemed to feel that the worlds financial system has become computerised to the point where human involvement in the descision making process is now minimal and, more frighteningly perhaps, there is actually no-one who understands or is in controll of it any more. (The same worries were of course bandied about re defense systems a decade or so ago, and it does do an author a bit of good to generate a bit of fear/interest in the subject area of his forthcoming book prior to it's release, but all the same .....)

peter wrote:The article I read was by an author called Robert Harris (of 'Fatherland' fame) who, while researching matereal for his new book on the financial wizards at the top end of Wall st, discovered that maths and physics graduates are being employed (rather than traditional economists etc) to produce ever more and more predictive tools that can forcast human behavior to the nth degree. He seemed to feel that the worlds financial system has become computerised to the point where human involvement in the descision making process is now minimal and, more frighteningly perhaps, there is actually no-one who understands or is in controll of it any more. (The same worries were of course bandied about re defense systems a decade or so ago, and it does do an author a bit of good to generate a bit of fear/interest in the subject area of his forthcoming book prior to it's release, but all the same .....)

there's a non-fiction book on that released a few months ago...don't recall the title, but it was talked about on cnn...the general opinion was he had the facts right, [which are the same as Harris says] but went over the top into conspiracy. What is true is that something like 70% daily trades take place without human involvement.

On self aware computers...we can't really know, of course. But one issue in it is: can awareness come/arise simply from speed and power? Or does it depend on the physical architecture? If it does depend on architecture, is there only one "form" that leads to it? Or a few? Or many? And does a different structure mean a fundamentally different variety of self-awareness? Would we be able to understand/communicate with it? Might there be some form that would make the Turing pass/fail irrelevant, cuz the test is Not Applicable?

Vraith wrote:On self aware computers...we can't really know, of course. But one issue in it is: can awareness come/arise simply from speed and power? Or does it depend on the physical architecture? If it does depend on architecture, is there only one "form" that leads to it? Or a few? Or many? And does a different structure mean a fundamentally different variety of self-awareness? Would we be able to understand/communicate with it? Might there be some form that would make the Turing pass/fail irrelevant, cuz the test is Not Applicable?

I don't think anyone else has ever asked those questions before. I know I have never thought of them.

I suspect that if its awareness takes a form different than ours and, from its point of view, its attempts to communicate with us fail then at some point it will quit trying.

The other questions/comments about different forms of awareness would need to go to The Close.

Hahaha, I'm pretty sure that those questions have been asked. They're good ones though.

I think (for no discernible reason) that intelligence is a function of neural connections. I don't think it matters whether those connections are protein or silicon.

And although we're still not quite close to AI, I do think that we're pretty unprepared for it.

As for your daughter, I think she probably won't need a flash drive either...computing is moving into the "cloud" at a rapid pace. All you'll need is something with a connection, and every program you want to use, and all your personal data will be available to you anywhere in the world.

--A

Avatar wrote:As for your daughter, I think she probably won't need a flash drive either...computing is moving into the "cloud" at a rapid pace. All you'll need is something with a connection, and every program you want to use, and all your personal data will be available to you anywhere in the world.

--A

As well as to anyone else who can hack their way into the cloud. I know that clouding is all the rage but--forgive my Ludditeness here--I distrust it because it is too open.

Avatar wrote: I think (for no discernible reason) that intelligence is a function of neural connections. I don't think it matters whether those connections are protein or silicon.

And although we're still not quite close to AI, I do think that we're pretty unprepared for it.


--A

I suspect that is a key...which is why I brought up architecture instead of just size/speed. [though I saw not long ago that they have made a transistor with graphene that is 1 nanometer...1/40th the current best commercial pentium size, IIRC...which is @ 10 years of so more of Moore staying true...and if I did my math correctly, cell phones would have 64 GIG processing speed.] Our brains are both massively multi-connected and massively parallel. Also, the neurons, though sort of on/off like transistors, ALSO do other things that transistors don't, they have internal states/voltage potentials...they also contain some portion of memory in themselves, not at a distant "memory" place that they have to address/search the index.
And for transistors, how fast they switch on/off shows nothing except how busy the system is. For neurons different firing rates often means nothing about how busy, it means it is perfoming a completely different kind of task.
So the neuron, unlike the transistor, is also multi-function.
FWIW, I think it would be far more interesting to find/make/discover a completely different kind of intelligence than one, relatively speaking, "in our own image."

And Hashi...yea, the cloud seems to me a digital nuclear holocaust waiting to happen.

I was expressing similar reservations about the 'security' of the Cloud to my daughter the other day and her response was that the kind of stuff one would store/acess from the cloud would be very trivial in the main such as music lists, films, photo's etc and so security was not quite as major an issue as it might at first seem. (May or may not be true - we'll have to wait and see, and she is always the eternal optimist to my perenial pesemist.)

re the self-awareness point - is not the very term defined by what we ourselves percieve it to be, and therefore anything that is not self-awareness of the type that we posess......is not self-awareness at all!

I feel the problem's less the data stored remotely, but the fact that people are pushing for us to have the applications we run stored there, too, so if the "cloud" does go down our devices will be nothing more than a fancy paperweight.

Murrin wrote:I feel the problem's less the data stored remotely, but the fact that people are pushing for us to have the applications we run stored there, too, so if the "cloud" does go down our devices will be nothing more than a fancy paperweight.

I'm guessing reliability of "the Cloud" would be an issue however I can imagine several ways it should be supported with redundancies that would make a fail nearly imperceptible. Not impossible that it could fail just that restoration would occur at the millisecond level.

Just typing millisecond, twice, has given me a headache.

But regardless of the datastore's reliability, you're still at the mercy of your own connectivity. I'm thinking of things like the Google cloud laptop, which was pretty much just hardware with a web browser.

Murrin wrote:But regardless of the datastore's reliability, you're still at the mercy of your own connectivity. I'm thinking of things like the Google cloud laptop, which was pretty much just hardware with a web browser.

Well connectivity is relative. There are an amazing number of devices for such things.
Heck even a micromodem brain implant could lose connectivity for some reason.
Fly by wire controls in use by today's military fighters still worries me. Imagine a full systems failure at 20,000 feet.

peter wrote:re the self-awareness point - is not the very term defined by what we ourselves percieve it to be, and therefore anything that is not self-awareness of the type that we posess......is not self-awareness at all!

But what if our self-awareness is inherently flawed and/or limited? What we think of as "self-awareness" might be a conclusion based on incorrect or insufficient evidence.

Murrin wrote:I feel the problem's less the data stored remotely, but the fact that people are pushing for us to have the applications we run stored there, too, so if the "cloud" does go down our devices will be nothing more than a fancy paperweight.

Agreed. And it's really only viable for people who have reliable high-speed broadband connections.

--A

Hashi Lebwohl wrote:

peter wrote:re the self-awareness point - is not the very term defined by what we ourselves percieve it to be, and therefore anything that is not self-awareness of the type that we posess......is not self-awareness at all!

But what if our self-awareness is inherently flawed and/or limited? What we think of as "self-awareness" might be a conclusion based on incorrect or insufficient evidence.

Yes - this is true even of our own self-awarenesss in that it appears that the land of our unconcious is apperntly forever closed to our 'aware' minds. Thus how much, even of our own selves, are we truly cognisent of what we are.

(Signing out for a week or so - off to London and Istanbul (not Constantinople!) for a few days )

I used to work at a processor manufacturing company and remember when the 3d architecture break through extended moores law. People were pretty pleased. As far as what comes if it extends, does this include switching to quantum computing if that ends up working out? It's not the same law, but it is an exponential increase curve that could continue on.

I don't think we in our 40s will really see the 'robot revolution', though. It would be an interesting time for sure, but I think it is too far away for our lifetime.

On what this self-awareness might look like, who knows. We assume it will look like ours, but I think that's hubris. My wild theory is that it will take the form more like an insect colony with the hive mind sort of thing. I guess that Star Trek outlined what that society looks like now that I think about it. Of course, their portrayal was a bit extreme.

The changes that I think we can see in *our* lifetime assuming another 40 years (at least for me) are smaller, I think. What comes to mind immediately is nanotechnology combined with computing. This is an emerging area and I could see some uses inside bodies in our lifetime for limited medical stuff, maybe to 'wire' us into a network, and of course for building of smart objects all around us.

For medical, I can't see a widespread use of this technology to dramatically increase lifespan because the world is already overpopulated. We don't need everyone living to 150 years. But, I see people who can really pay having access to this. In our lifetime, I think it will only be the start of such things, though. After we're dead and gone, maybe the grandchildren of our kids will have these options if they are very rich!

For hooking us into a network, I think this is something corporations and governments would love to do! It could track us all the time. Our access to things can be very easily controlled since access would be granted through a personal implanted device rather than a terminal somewhere. We could become walking, talking receptacles for product consumption to even a greater extent than we are now. And, someone will always know what we're up to and where we are. I think this is something that would only come at the end of our lifetimes.

And smart objects, I guess those could come in the form of clothes that changes depending on weather and temperature and so on. Buildings that adjusts for conditions. Lighting. I don't really know. I've read a little about the applications for these things, but not much, so maybe my imagination is limited on these. Again, not till the end of our lifetime for these to be common, though.

I forgot warfare! I don't even want to think of these.

If the law fails, then Apple's stock price will go down again. They can't sell us a new device every two years anymore. On the other hand, there might be a slight decrease in toxic electronics waste being shipped to afrika for dumping. All in all, I think it won't be a big deal at all if it fails except to some people's stock portfolio.

Ananda wrote:For hooking us into a network, I think this is something corporations and governments would love to do! It could track us all the time. Our access to things can be very easily controlled since access would be granted through a personal implanted device rather than a terminal somewhere. We could become walking, talking receptacles for product consumption to even a greater extent than we are now. And, someone will always know what we're up to and where we are. I think this is something that would only come at the end of our lifetimes.

Tangentially related, they're making big steps forward in the realm of thought-controlled prosthesis technology right now. They're actually at the stage of trying to move from wired to wireless brain implants - being able to control machines and computers with thought wirelessly. That's something that has the potential to leave the medical field and turn into something big 40 or 50 years down the line.