Moore's Law and the power of exponential advancement

[Zog]

I'd speculate that in the not-so-distant future, people will not have children anywhere near the rate they do today and will live and considerably longer than they do at present - thus, the incentive to have children, due to say a need to live on, will be greatly diminished.



Absolutely agree..

I have 2, one sitting right next to me. .

What is the best place to buy a 3D printer, and what does a reasonable good one cost?

 

[gourimoko]

What is the best place to buy a 3D printer, and what does a reasonable good one cost?

Amazon.

As far as cost, it depends on your technical/building ability.. If you're comfortable assembling a printer and maintaining one, then you simply cannot beat the QIDI Tech I which goes for around $599 but is the equivalent of a $999 3D printer (dual extrusion, replicator compatible, etc)..

You will need to do some tinkering, and some settings adjustments; but the QIDI consistently outproduces my FlashForge Creator Pro which is $300 more expensive. The only thing is that the FlashForge is easier to configure... but the QIDI, as generic and cheap as it is, is just as good of a printer all things considered.

I will say that having maintained a QIDI (and a RepRap long ago), I'm much much more comfortable with the idea of building a 3D printer from scratch -- and even my wife is comfortable maintaining the damn thing.. She did a nozzle repair on the printer which was rather unexpected.

They seem like black boxes, and some of them are, but that's why I like the generic QIDI as a printer that's good for basic prototyping.

With that said, the Ultimaker is up next on my to-get list once I move into a bigger place with more space.. That would surely retire the printers I'm using now for certain.

 

[Adam]

I found this very enlightening although still a long way to go

 

[The Oi]

Amazon.

As far as cost, it depends on your technical/building ability.. If you're comfortable assembling a printer and maintaining one, then you simply cannot beat the QIDI Tech I which goes for around $599 but is the equivalent of a $999 3D printer (dual extrusion, replicator compatible, etc)..

You will need to do some tinkering, and some settings adjustments; but the QIDI consistently outproduces my FlashForge Creator Pro which is $300 more expensive. The only thing is that the FlashForge is easier to configure... but the QIDI, as generic and cheap as it is, is just as good of a printer all things considered.

I will say that having maintained a QIDI (and a RepRap long ago), I'm much much more comfortable with the idea of building a 3D printer from scratch -- and even my wife is comfortable maintaining the damn thing.. She did a nozzle repair on the printer which was rather unexpected.

They seem like black boxes, and some of them are, but that's why I like the generic QIDI as a printer that's good for basic prototyping.

With that said, the Ultimaker is up next on my to-get list once I move into a bigger place with more space.. That would surely retire the printers I'm using now for certain.

What kinds of things do you create most often using your 3D printer? What kinds of things could I as a regular dude print?

 

[The Human Q-Tip]

What makes you say "many centuries in the future?"

Because my life experience has been that these predictions of future changes coming on us always underestimate the amount of time required for the transition. I think I've mentioned that before, so I won't go into the specifics.

The biggest issue to me is the gap between the technological ability to do something, and the practicalities of having that technology actually be reduced to percolating completely down through all levels of society. It's the difference between a World's Fair "Home of the Future" exhibit, showing everything that is then technologically possible, and everyone actually having such a home.

Sure, it will be possible for the very wealthiest in society to live a life that is highly automated. But the world has nearly 7.5 billion people, many of whom live in abject poverty in incredibly remote places. Bringing that level of automation to them is simply a gargantuan task.

As I see it Q-Tip, in the not-so-distant future, perhaps in our lifetimes, we will very likely have a complete artificial brain; as in, a conscious, sentient, thinking machine modeled after the human brain via a complete simulation of the human neural network.

I don't think that's true, but let's assume it is -- that's kind of my point. Great -- we've got this brain. Now what do we do with it? I mean, we've already got 7.5 billion of the things already, and they agree on very little. So we've just invented something redundant. Why would the great mass of actual humanity be willing to agree with/trust such things any more than they agree with/trust other humans with brains right now? We're a fractious lot, and always have been.

Once we get to the point where such simulations can do complex tasks, like my job as an analyst/programmer, or say your job as a lawyer, then yes, the point where automation begins to reshape and ultimately obsolete the market economy will be here.

Okay, just to break this down, you started with "perhaps in our lifetimes" we'll see a brain. But from that point of technological creation, you then have automation "beginning" to reshape and "ultimately obsolete" the market economy. But you've got to get from the point of having a brain (and big deal, we already have them), to completely automating every bit of work currently performed by humans. the capital investment in machinery alone would be beyond astronomical. Building wells, roads, and schools, farming, ranching, home building, cleaning, cooking, repairing...there is a virtually endless list of tiny little tasks essential for human life, and having all of them automated likely would never make economic sense. That ignores the fact that humans may simply prefer, purely for aesthetic reasons, to have a great many tasks/services performed by other humans.

Suppose you want to see a move, or play, or go to a concert, or watch a sporting event. Are we all going to want to see robots doing such things, or real live performers getting paid for their work?

Or at least, go back to the point I made initially. Is it really too much to think that it will take "many centuries" before everyone in subSaharan Africa can lay back in their beautiful home, with limitless food, medical care, transportation, without actually working? I think with "many centuries", I'm being optimistic. The cultural hurdles/resistance alone before you can even begin to do that are going to be huge.

The near-future technology that's on the foreseeable horizon puts us fairly close to this point where we reach a technological singularity; at which time, technological advancements will vastly increase in pace, quite obviously due to the much faster cognitive speed at which artificial brains can operate.

Again, you're talking about the discovering/devising the necessary technology. I'm pointing out that the implementation of all that technology and supporting infrastructure for humanity is a massively more time and resource consuming project. Frankly, we may not even have the resources to make that level of investment in infrastructure feasible or worthwhile at any point.

 

[KI4MVP]

Because my life experience has been that these predictions of future changes coming on us always underestimate the amount of time required for the transition. I think I've mentioned that before, so I won't go into the specifics.

The biggest issue to me is the gap between the technological ability to do something, and the practicalities of having that technology actually be reduced to percolating completely down through all levels of society. It's the difference between a World's Fair "Home of the Future" exhibit, showing everything that is then technologically possible, and everyone actually having such a home.

Sure, it will be possible for the very wealthiest in society to live a life that is highly automated. But the world has nearly 7.5 billion people, many of whom live in abject poverty in incredibly remote places. Bringing that level of automation to them is simply a gargantuan task.

Look at how fast technology is being adopted. The first modern smartphone, the iPhone, was released less than 10 years ago. There are over 2 billion smartphones in use today (about the same as the number of personal computers). 3.5 billion people have internet access. nearly 2 billion people use Facebook (which was founded just 12 years ago).

And the issue really isn't how long would it take to adopt new technology worldwide, it's how long would it take to adopt new technology nationwide.

I don't think that's true, but let's assume it is -- that's kind of my point. Great -- we've got this brain. Now what do we do with it? I mean, we've already got 7.5 billion of the things already, and they agree on very little. So we've just invented something redundant. Why would the great mass of actual humanity be willing to agree with/trust such things any more than they agree with/trust other humans with brains right now? We're a fractious lot, and always have been

Once we can build a computer brain that works as well as a human brain, we can build ones that works exponentially better and exponentially better faster by following Moore's Law.

Just look at what's happened with Watson. It's gone from first competing in Jeopardy 6 years ago to becoming a service any business can use, including building smartphone client/server apps that utilize the power of Watson.

 

[Otsego Mar]

What kinds of things do you create most often using your 3D printer? What kinds of things could I as a regular dude print?

All the dildos your wife could neeed. See, technology at work, freeing you up to do the things you really love, like posting on RCF.

 

[FiveThous]

post on RCF

Some of us don't need automation for that.

 

[KI4MVP]

I found this very enlightening although still a long way to go

there should eventually be a few more steps to that. Curing disease (including the largely ignored disease called aging) would dramatically increase average expected lifespan. It would still lead to eventual population stabilization as long as the birth rate stabilizes at or below 2 children per woman.

 

[The Human Q-Tip]

Look at how fast technology is being adopted. The first modern smartphone, the iPhone, was released less than 10 years ago. There are over 2 billion smartphones in use today (about the same as the number of personal computers). 3.5 billion people have internet access. nearly 2 billion people use Facebook (which was founded just 12 years ago).

That is different from building an automated infrastructure for every task now performed by humans. It is orders of magnitude different.

Once we can build a computer brain that works as well as a human brain, we can build ones that works exponentially better and exponentially better faster by following Moore's Law.

Moore's Law isn't an actual law.

 

[KI4MVP]

That is different from building an automated infrastructure for every task now performed by humans. It is orders of magnitude different.

It's already in progress. For example, self driving cars are a reality on the road today (Tesla AutoPilot 2.0).

Moore's Law isn't an actual law.

I know that, that doesn't change what I said. Once we can build it, we can scale it up exponentially. We can also make it more reliable than the human brain (cells don't die in a computer).

 

[The Human Q-Tip]

It's already in progress. For example, self driving cars are a reality on the road today (Tesla AutoPilot 2.0).

To what end? Again, you are focusing too much on the chip/computer end of things. I am thinking more along the lines of the billions of tons of goods/raw materials produced and shipped from all over the place. Putting a plane on autopilot is an incredibly less daunting task than actually having planes designed, marketed, and produced (including acquisition of raw materials) without any human input at all.

I know that, that doesn't change what I said. Once we can build it, we can scale it up exponentially. We can also make it more reliable than the human brain (cells don't die in a computer).

But we may not be able to scale it up exponentially. Moore's law describes what has happened in the past. It does not dictate the future, and the initial growth rate associated with Moore's Law has in fact slowed already. Moore himself predicted it would die out completely in less than ten years -- which is long before anyone thinks we'd succeed in actually developing the equivalent of a human brain.

There also is the issue of how much trust humans are willing to put into machine "brains", assuming we reach that point. For certain area of research, sure. But human existence consists of a hell of a lot more than research, and it is entirely possible that a "better" set of cybernetic brains will perform in a manner completely unappealing to humans.

 

[KI4MVP]

To what end? Again, you are focusing too much on the chip/computer end of things. I am thinking more along the lines of the billions of tons of goods/raw materials produced and shipped from all over the place. Putting a plane on autopilot is an incredibly less daunting task than actually having planes designed, marketed, and produced (including acquisition of raw materials) without any human input at all.

Perhaps we're looking at different ends.

We are much further from having programmers, designers, etc. being automated. Other industries are going to get massively automated. Foxcon has already announced a 3 step plan to completely automate their manufacturing process over the next several years.

I'm not expecting everything to be replaced, I am expecting enough to get replaced that we have to dramtically rethink society. Even without automation, aspects of our increasingly service based society are being replaced with self service.

And I would hope a major goal of computer brains is to cure disease (including aging). There is far too much data for any one person to understand and process. Computers are getting better and better at processing massive amounts of data. You don't have to blindly trust the results for something like that, you do like we do now and test the results.

 

[The Human Q-Tip]

Perhaps we're looking at different ends.

We are much further from having programmers, designers, etc. being automated. Other industries are going to get massively automated. Foxcon has already announced a 3 step plan to completely automate their manufacturing process over the next several years.

I'm not expecting everything to be replaced, I am expecting enough to get replaced that we have to dramtically rethink society. Even without automation, aspects of our increasingly service based society are being replaced with self service.

My point was in the context of @gourimoko saying no more free market economy because of automation. That essentially requires that human labor/input is of no value. To reach that point, I think, will take an incredibly long time, with so many other possibilities that projecting that far out is just not possible. The thing is that I think the poor will always be engaging in some form of value creating work -- even among themselves -- because the investment cost to provide them a fully-automated everything simply will not be worth it to those with the assets.

And I would hope a major goal of computer brains is to cure disease (including aging). There is far too much data for any one person to understand and process. Computers are getting better and better at processing massive amounts of data. You don't have to blindly trust the results for something like that, you do like we do now and test the results.

No argument there.

 

[gourimoko]

Because my life experience has been that these predictions of future changes coming on us always underestimate the amount of time required for the transition. I think I've mentioned that before, so I won't go into the specifics.

Right, which while surely reasonable, isn't appreciative of what exponential growth really entails. While, during your lifetime, you've seen what you perceive as moderate growth; your perception of that growth is preventing you from realizing that there is a snowball effect in play where continual technological advancements will come at an ever increasingly faster pace. Such that, the gaps between technological jumps in your lifetime will be tremendously smaller compared to that of your children and grandchildren.

This is the whole point behind Moore's Law..

The biggest issue to me is the gap between the technological ability to do something, and the practicalities of having that technology actually be reduced to percolating completely down through all levels of society. It's the difference between a World's Fair "Home of the Future" exhibit, showing everything that is then technologically possible, and everyone actually having such a home.

I would agree with you for certain technologies; but with the two technologies I've described to you earlier; they are specifically designed for wide-scale use. I'm not sure how you would use this analogy with reference to the hypothetical I described above?

Sure, it will be possible for the very wealthiest in society to live a life that is highly automated.

Why the wealthiest? Again, I've gone over why this would cease to make sense once 3D printing, robotics, and AI reach such a point as they can operate one another. Hence the whole concept behind a "technological singularity."

But the world has nearly 7.5 billion people, many of whom live in abject poverty in incredibly remote places. Bringing that level of automation to them is simply a gargantuan task.

Q-Tip, that's the whole point....

It's a gargantuan task now.

But you're not factoring in what happens when we can literally print/assemble autonomous machines that can raise people out of poverty, permanently. When we can grow food without animals or excess land area or irrigation... I mean, I are you taking into consideration what such advances in artificial intelligence and robotics would really entail?

People living in poverty could still buy (or build) a 3D printer; print and assemble an android, and in a matter of years, after the process has repeated itself enough times; everyone has an android and a 3D printer -- because the devices are essentially self-replicating.

I don't think that's true, but let's assume it is -- that's kind of my point. Great -- we've got this brain. Now what do we do with it?

Really?

Q-Tip, with the ability to simulate an artificial brain in a virtual machine then we could scale complex tasks across machines -- meaning that areas like medicine and science can be handed over to cloud computing services. Instead of assembling a team of 25 programmers, you just log some time in the AI cloud and it handles everything for you.

This... literally changes everything. It means you and I are out of a job.

I mean, we've already got 7.5 billion of the things already, and they agree on very little.

The purpose of building such a neural network is to understand human creative and cognitive ability. If you were to scale this up solely for computational purposes then you wouldn't simultaneously include areas of the brain dealing with emotion which largely lead to these disagreements to begin with.

So we've just invented something redundant.

I get what you're trying to say, but I think you might misunderstand -- you're not simply reproducing the human brain for the sake of it, you're building a simulation of it in a virtual machine so that you can scale cognitive ability and problem solving onto standing computing platforms.

This means that you could have a human-like cognitive capability scaled over silicon. That's not redundant -- that's a game changer... in every way imaginable.

Why would the great mass of actual humanity be willing to agree with/trust such things any more than they agree with/trust other humans with brains right now? We're a fractious lot, and always have been.

I'm not arguing that AI govern humanity -- I'm arguing that AI can serve humanity. You raise an interesting point, but I think once we grasp that human consciousness can potentially exist in a purely artificial space; then we'll likely stop thinking like this anyway, i.e. "such things" would be our wives, husbands, and children.

Okay, just to break this down, you started with "perhaps in our lifetimes" we'll see a brain.

Yes, I think so.

But from that point of technological creation, you then have automation "beginning" to reshape and "ultimately obsolete" the market economy.

Yes.

But you've got to get from the point of having a brain (and big deal, we already have them),

Re-read above... Having an artificial neural network modeled after the human brain literally IS the point where you've reached the technological singularity -- again, because problem solving at that point is now an issue of potentiality, manufacturing capability, and time rather than education, socioeconomics, human labor, and biological brainpower.

to completely automating every bit of work currently performed by humans.

Not every bit of work.. Again, if you go back I stated that there would still be things for humans to do -- but that there would be no structure in place for an economic system to function.

the capital investment in machinery alone would be beyond astronomical.

No.. it wouldn't.. You're not thinking about what we're saying...

3D printing and androids make the cost of building and repairing machines and structures dirt cheap, if not "free." Where does the astronomical cost come from?

Q-Tip, there is no labor. The only thing that is required is raw mineral resource which can be mined by machines working 24/7 for free, who self-replicate, and repair themselves, and print new machines and androids to scale for projects that are larger or smaller....

Where is the astronomical cost?

Building wells,

Androids.

roads,

Androids.

and schools, farming, ranching, home building, cleaning, cooking, repairing

Androids, androids, androids...

I don't think you really get it.. Why would you need a person to farm if you have an autonomous android that knows 10x more than anyone you could hire, can work 10x faster, is 10x stronger -- needs no salary, food, water, shelter.. The only thing these things require is power which can be freely extracted from the sun?

...there is a virtually endless list of tiny little tasks essential for human life, and having all of them automated likely would never make economic sense.

You're right, it makes zero economic sense, that's why the economy would cease to exist.

Do you refuse to use an android because it doesn't make sense on the larger scale, or do you use one because it's the best thing for you individually?

That ignores the fact that humans may simply prefer, purely for aesthetic reasons, to have a great many tasks/services performed by other humans.

I agree with this, in the short-term.

But in the long-term, people will be marrying these androids and AI... I think you're thinking that things will be as they are, without appreciating the societal effect of having the equivalent of a human intelligence inside of an anthropomorphic robot that's designed to cater to your every need.

Suppose you want to see a move, or play, or go to a concert, or watch a sporting event. Are we all going to want to see robots doing such things, or real live performers getting paid for their work?

Rogue One Box Office: $829.1 million

Or at least, go back to the point I made initially. Is it really too much to think that it will take "many centuries" before everyone in subSaharan Africa can lay back in their beautiful home, with limitless food, medical care, transportation, without actually working?

Yes, it makes no sense as to why it would take so long.

Q-Tip, once it's developed here, it'll be used there, they don't need to go through all the steps we did to get to the same point.

It didn't take many centuries for these folks to have cable TV, cellphones, and the internet.

I think with "many centuries", I'm being optimistic. The cultural hurdles/resistance alone before you can even begin to do that are going to be huge.

I'm not sure what cultural hurdles you're referring to or who would resist such forms of automation? Moreover, how do you stop someone from buying the first lot of self-replicating android systems? You protest them? I'm not sure how that works in the long-term?

Again, you're talking about the discovering/devising the necessary technology. I'm pointing out that the implementation of all that technology and supporting infrastructure for humanity is a massively more time and resource consuming project. Frankly, we may not even have the resources to make that level of investment in infrastructure feasible or worthwhile at any point.

I don't see why you're saying what you're saying though.. You seem to be thinking that it's a monumental task, when it's not.

The whole point is that you build the initial system which is self-replicating and self-growing. At that point, human interaction is minimal if not simply unnecessary.

Where is the monumental challenge?