Ask Ray | The future of Moore’s law
May 28, 2013 by Ray Kurzweil
Dear readers,
A recent article in EE Times discusses the future of Moore’s law: “Broadcom: Time to prepare for the end of Moore’s Law.” This comes up innumerable times. People assume that Moore’s law is synonymous with my law of accelerating returns, which it is not.
Moore’s law was the fifth, not the first, paradigm to bring exponential growth to the price-performance of computing. The law of accelerating returns does not pertain to feature size or clock speed but to calculations per second per constant dollar.
Three dimensional computing will keep that exponential trend going for many more decades. I’ve long said that the paradigm represented by Moore’s law (shrinking transistor sizes on a flat integrated circuit) will come to an end in the early 2020s.
— Ray Kurzweil
comments 72
by gs4mac
Let’s face it, the real dark side of the law of accelerating returns is accelerated obsolescence:
http://vimeo.com/14388358
by JustDavid
The bright side of the dark side is that obsolete tech keeps poor people and children in touch. When I read of new gadgets, I think ‘Goody! I’ll be finding these in the dumpster in a couple years.’
by Durk
I am a freelance portrait artist whose annual income is probably less than $7000. Yet I own at least six computers, three of which are multiprocessor base. Trickle down or not I have benefited from advances in technology.
by Ed Pell
The price points have to do with what is profitable for the seller and will have enough willing buyers. If you just want to make call you can buy a $30 cell phone.
by GatorALLin
So why are cell phones still at $500 value every single year? Sure the features go up with the number of apps available, but that is software, not hardware. Sure you get a $200 phone, but you sign up for a $99/year family plan for 2 years, so that is $2,400 and thus you really pay $600 for your phone every 2 years. Are we as customers just getting sucked into a marketing gimmick and thus cell phones are in a bubble of some sort……??
by DAG
Phones’ computational abilities are drastically increasing. Today I can tell my cell phone to do an internet search, find and start streaming an HD video in seconds, access all of my contacts, calendar, and emails instantly and seamlessly, search a world-wide map quickly and get moment-by-moment navigation on a whim (for driving, bicycling, or walking)… I can search and instantly listen to more HD music than I’ll ever have time to listen to… and the list goes on. Most of this was unimaginable to the general public just a few short years ago. Ten years from now our phones will probably be doing things we can hardly imagine now.
by Gabor
Exactly right!! Smartphone prices are misleading. They are not phones anymore but computers that costed thousands of dollars 5 years ago. And 5 years from now they will easily surpass (and likely replace) the best desktops of today and still be in the $5-600 price range. They do get more hardware in them, too, all the time (although it’s not always obvious but transistors are getting smaller and much larger in numbers within).
Non-smart “Cell Phones” are actually free (or almost free) now at major providers as long as you pay for a minimum plan.
by pt
Graphene cores + highly parallel processing would seem to take it beyond 2020, though I suppose once quantum computers start getting to extreme numbers of qubits, it’s no contest…so it may just be that quantum pcs will be improving faster than graphene processors will be able to compete with. That whole doubling processing power for every qubit added thing is a pretty absurd advantage.
by Jack Reeve
At the molecular level, digital is still analog. There is no end to accelerating returns. I just love that we are here to see the new beginning. That it is unimaginable at only 40-50 years out is to me an irresistible concept. I await that golden day – the day after the day that each day’s technology extends my life by one day. That day, my friends, is a day of magic. Have a nice day.
by Jesus Christ
I doubt that any technology to extend human life will be available to just anybody, it will most likely be only available to the elite mega rich. The elite don’t want the average person living forever creating an even worse over population problem and draining ever dwindling resources.
by glen
What about the average elite or elite average person?
by dj
This is why we must change our mentality from the profit orientation that you see today to a system that considers all. A system that claims all of the earths resources a common heritage to all members of this planet. That is how the coming singularity will be the singularity that ray truly envisions.
by Positron
Kuzweil says that the rich get technologies when they don’t work. Initially life extension technologies will likely be out of reach for many but eventually it will be cheaply available to everyone.
http://arstechnica.com/business/2013/06/google-director-early-adoption-is-for-rich-chumps/
by Boy
what about special chips, in the cloud that can do some specifi kind of job very well, somehow the internet of special chips in cloud my become, new procesor?
by Marraco
I suspect that the most promising new paradigm is multivalued logic (google it). Think of polinary transistors instead of binary ones.
Is not a new concept, a trinary computer named Setun was made in the 50′s.
A 4 states transistor could emulate 2 binary ones, on the same space, without need of further miniaturization.
Multivalued logic is already working on MLC flash memory.
Instead of reducing the size of transistors, this new paradigm would progress by raising the number of states with each generation. it would go from binary to 4-ary, 8-ary, 16-ary…
by Jack Richardson
This is an example of how a new kind of software can provide the doubling that you would get from 3d hardware.
by Clark
Better to go directly to full analog for appropriate computations. See http://www.amazon.com/Analog-VLSI-Neural-Systems-Carver/dp/0201059924 for a good introduction to established state of the art.
by me
But fully analog requires new software.
A polinary, digital computer can be designed to be transparent to software. If a transistor is doing the work of 2 or more transistors, the software doesn’t need to know.
by Cybernettr
Reading the comments on the original article, it’s amazing how many “techie” types are unaware that the end of Moore’s Law doesn’t at all signal the end of the doubling of the price/performance of computer technology! Most people are shockingly clueless when it comes to the technological future.
Perhaps when Moore’s Law runs out, Kurzweil’s LOAR will take it’s place as the de facto term to describe the continuing exponential growth of information technology.
by Boy
how about power? in 3d chips?
by franco
What if there is not another paradigm after microchips? Do you really believe we can store a computer data and capabilities inside an atom? How exactly?
And do you really believe we can upload our minds inside a computer like “Tron” and still be alice and conscious? I tell you it’s just nonsense. Computers cannot become little as an atom nor they can become conscious. That’s why Singularity will not happen and next development of our species will be exploring paranormal powers and spirituality and religion.
by VovixLDR
Although I doubt if it makes sense to reply to your comment at all, but are you really so blind to the reality that you cannot imagine e. g. quantum computer but still believe in paranormality and “religion as development”?
> I tell you it’s just nonsense. Computers cannot become little as an atom nor they can become conscious. That’s why Singularity will not happen
BRILLIANT ARGUMENTATION! “It cannot be because it can never be!”
by Marraco
The brain is a computer, so we know that is possible
by Robert Duckmanton
A fallacious analogy. The brain is not a computer. It is an evolutionary device to maintain complex organisms in complex, constantly changing environments in optimal health. We know for a fact that it is the source of mind and consciousness, which so far we are not able to suppose in relation to computers, however fast.
by Marraco
The brain is a physical device, which processes information. So, is possible to make an intelligent physical device which processes information.
by Sett
Increasing computing power? Ridiculous!
Paranormal superpowers? Coming soon!
..right.
by Rok
What is the objective mathematical formula for law of acelerating return or are we changing over time what are we calculate, so the formula is ok?
by eldras
This is really helpful thanks.
“Anything that can be said, can be said clearly” Wiggensteion Tractatus. Proposition 4.116
by coolball
I think 3d computing’s ability to continue the doublings will depend on both ease of manufacturing ever more components without price escalating(probably will require more advanced nanotech manufacturing technologies), and the ability to keep energy and heat in check maybe requiring reversible computing
by Peter Kinnon
It is so sad that Ray Kurzweil and his cohorts should have such a comprehensive grasp of the exponentiality of technological development yet, trapped by our inherent anthropocentricity, remain completely blind to the essential nature of the process.
Viewed objectively, using understandings acquired across a wide section of the scientific spectrum, it becomes very clear that this process is merely a phase of the more fundamental “life” process and that it’s future direction lies not in enhancement of our particular biological species but rather in the establishment of the completely new inorganic phase of which our present Internet is a progenitor.
In recent years I have advanced this theme in my writings and while this realization is gaining traction it has yet to make a substantial impact on the main transhumanist community, who desperately cling to the time-honored notion that we little bipeds are somehow central to the functioning of this universe.
The basis for such thinking is expanded upon very informally in “The Goldilocks Effect: What Has Serendipity Ever Done For Us?” , a free download in e-book formats from the “Unusual Perspectives” website.
The more formal work on which I started “The Intricacy Generator: Pushing Chemistry Uphill”, may, owing to a recent diagnosis of advanced stage cancer, never see the light of day. So it may be that my previous writings, with all their inadequacies, will have to suffice. I have tried my best.
by melajara
I’m very sorry to read your last paragraph. Despite your last sentence, I would urge you to keep fighting.
With all the lab news we keep reading day after day on this website on encouraging results to blast cancer tumors in imaginative ways, I would advise you to look for volunteering in the kind of treatment most fit for your particular form of cancer. Maybe Ray or our benevolent editor could help with links here.
I will download your book now, thank you for sharing
by Editor
Really sorry to hear. I’m in touch with him.
by Jack Richardson
Ray can correct me if I’m wrong, but I believe his vision coincides with yours in the sense that he envisions the eventual migration of our minds into an inorganic environment that supports an indefinite existence.
by Peter Kinnon
Actually, Jack, I do not propose that human minds will migrate to the next phase of the “life” process. Rather that this new entity is evolving autonomously within collective human imagination. A very different concept.
This evolutionary process is not limited to technology, or even to biology but can be traced at least as far back as the evolution of the chemical elements in stars and supernovae.
by Jack Richardson
Possibly you have in mind a version of the Gaia notion of Earth as a conscious entity that is evolving from the life forms that it has generated up to this point. As humanity takes control of biological evolution in coming decades, Gaia may control the direction this takes. At the same time, AI will be advancing beyond the point where humans can keep up, but will Gaia be able to keep up? Or is Gaia a small part of a larger conscious solar system?
by eldras
Good effort Peter. Very hard to see with increasing perspectives on cosmic scales.
http://www.unusual-perspectives.net/goldilocks.html
by tedhowardnz
Hi Peter
I was given a 50% probability of living 5 months 3 years ago (advanced metastasised melanoma).
See my blog for what I did.
I am tumour free.
Short version, take high dose vitamin C – heaped teaspoon in glass of water twice a day, and 1/4 tsp in glass of water every waking hour in between; and remove all animal protein from your diet. Not easy, and I am still here.
My medical records at http://www.tedhowardnz.com/about, and what I did diet-wise at http://www.tedhowardnz.com/cancer
Best of luck.
I can be done.
It isn’t easy.
Give it your best shot.
Ted
by r. m. dobbs
Peter, you’re too novel not to survive somehow.
by glen
“Could the internet and technology be the end of us? Yes, it could.”
How?
by Kristof77
??? not sure where your trying to go with the comment, but technology will eventually allow mankind to take control of our evolution, so in that sense, yeah it could be the end, althought im thinking your just trying stir up the bees.
by Gabor
I know I’m not Ray, but the answer to your question is so easy if a little scary that I cannot resist.
Yes, “the internet and technology” will be the end of humanity as we know it. The only question is “when” and not “if”. To be more precise, it is a continuous (exponential) process and we are by far not the same humanity as we were a 1000 years ago. Yes we still have the same barbaric urges but we are controlling them better (for the most part) thanks to technology and information abundance.
1. “Internet” means information availability – in the next few decades, the internet will evolve into a network of seamless database to which every one of our minds will be able to add and subtract information on a real time bases as well as most of everything (possibly to the atomic level) will be connected and can be controlled. Telepathy, telekinesis, teleportation anybody?!?
2. Technology means processing, putting order to chaos and do it with an exponentially evolving efficiency. The use of energy to create more efficiency (order) is growing exponentially. For example think of creating energy from oil and solar means. Oil is stored solar energy that took millennia to “create”, but solar energy is converted almost immediately with solar panels.
Technology is allowing us to process matter and energy on ever smaller scales until we reach the atomic or even subatomic levels. Maybe Star Trek’s replicators and Holodeck that’s every bit as real as reality but is programmable come to my mind.
The internet and technology will be the end of humanity when we will progress them to the point where they allow us to change our own genes freely and eventually allow our minds to transcend completely the substrate dependence. In my opinion it is certainly in reach within this century but the crucial period will be the next 20-30 years and as optimist as I am if asked to bet on our survival, I would definitely get a short position (would bet against our survival).
It’s an arms race between technological evolution and the controlling unbelievably outdated “software” that is our brains. I’m not talking about war between machines and humans, technology is not good or bad. I rather talk about a fatal error in controlling powerful technology due to outdated human brains. The most important research (life line) is to upgrade the good old software. The brain projects (both US and European) and efforts of brain machine interfaces as well as developing the all “powerful” AGI (Artificial General Intelligence) should get the highest priority, even over military and cancer research as the later ones can only delay doom’s day but not prevent it.
I did say I am an optimist though but people will need to understand that as much as new technology is making our lives better and easier. The next few decades we will have to surround most if not all of our freedom perhaps in the form of 24 hour surveillance of every free roaming individual with the possibility of stopping them in their tracks. I know it sounds bad but this is our only chance until we can upgrade. How many of your neighbors would you trust with a RED BUTTON that would annihilate Earth if pushed?!? Would you trust yourself?!?… say when you are under the influence of drugs or alcohol or depression??? Well in a couple of decades not just your neighbors in the next house but “neighbors” in other continents will be able to get their hands on those “red buttons” as easily as they can get a cell phone today! Our ultimate defense is to make everybody smarter, much smarter!
by JC
Great post. Hope to read many more. But you don’t really mean ‘smarter’ do you….a smarter sociopath is just a more effective one. Are you hesitant to say mind control? or thought suppression? The ‘control of barbaric urges’ will take the most severe methods. Don’t you think?
by Vin
Maybe Hugo de Garis’ thesis. Cyborg (necessarily weighed down by evolutionary baggage to some degree) Verses Artillect (not, to any or negligible degree)? Depending on say (1) Artillect survival is threatened and they have a survival motive.above xenocide (2) There is a critical competition for resource. (3) Morality or empathy is exclusively inversely proportional to intelligence.
by Tom
competition, of course.
By creating what out-competes us.
Nothing new here.
Billions of species have been doing this for billions of years by creating the mutated offspring that eventually replaces them.
Why should we be any different?
HOW we will be different is that we may create a new type of life.
But, unless we manage to become a part of this new type of life, we will be doomed to going the way of all species.
by tim the realist
I really only understand the / sec part on the vertical axis scale. Please clarify what is considerered a “calculation” on the gragh. Is there a standard calculation used across the entire time frame for the data points? If so, and it’s a simple calculation, then it’s mostly a hardware capabilities graph. If it’s based on a complicated enough function, then this graph would also take into account software advances.
Also, is it $1000 from 1900 inflation adjusted to today? Or the other direction?
It’s also only the growth of electrical computing – no abacus calaculating or people calculating by using paper and pencil, lookup charts,or slide rules are included.
Maybe i’m over analyzing it?
by Dr. Richard
So far this is pure speculation. Man-computer fusion, 11D computing….all science fiction or worse, fairy tales. Nothing much has happened yet at all beyond the Integrated circuit….lots of ideas, nothing else. No real super computers except variations on the same theme. Lots of talk, hot air….we are all going to be super human, computers will be our Gods….mostly bullshit so far.
by Dr. Bloggs
Yes, the predictions from these idiots are preposterous! Why, Kurzweil predicted in 1986 that the Soviet Union would collapse due to electronic dissemination of information, that there would be billions (yes, with a B) of people connected daily to the Internet by 2010, and that a majority of people would connect to the Internet many times daily via devices that have no wires. As if!!! These people live in a dreamworld!
Sent from my iPhone
by Paul Denton
Looks like you’re behind the curve, Dr. Richard. With the recent roll out of D-Wave’s computers, it is expected that biotechnology is set for a paradigm shift, as protein folding becomes millions of times quicker. What about IBM’s Watson? Are you not impressed that computers are now better than humans at answering natural language questions?
I look around me and I see massive accelerating progress in computing and biotechnology. There is no reason to assume that thousands of years of accelerating technological progress are going to grind to a halt because the future hasn’t happened already. All predictions are science fiction. But some predictions come true.
by A J
Thanks for mentioning D-Wave.
Too bad QC is not mentioned here (in the Ask Ray).
by Ron
Dr.(?) Richard, maybe you are only 20 something but I can see huge changes in my lifetime. If you think I could have made a phone call anywhere I wanted from a small mobile computer while checking the news and weather when I was 25, you would be dreaming. I used the internet on a 300 bps modem and my 25 Mbps connection sure beats the hell out of that.
by pauldenice
In addition to Moore’s law on hardware componants it is also necessary to take into account the exponential progress made in software, in particular through knowledge networking and personnal accesibility to computer power which were only available to a few specialists in large business enterprises or university labs.
Software Creativity can no emerge from very small groups or even individuals: here is an example from a young Indian programmer Pranav Mistry
If you haven’t seen this video take the time to see it:
http://www.ted.com/talks/pranav_mistry_the_thrilling_potential_of_sixthsense_technology.html
by fore//
In the future the poorer will have a much harder time.
Already in our time, the wealthier have access to cutting edge treatments to cure and diagnose several diseases that will kill the poor. Special health clinics help the wealthy to live longer, like the ones Ray himself regularly visits.
NOW, i am not criticising the wealthy, i am just saying how things are and will continue to be. You want something better for yourself? You want to have a higher chance of living forever? Go and do some real hard and smart work, put together all the money you can, because sure as hell you’ll need it when the real technological revolutions dawns upon us in the next several decades.
Stop complaining and start acting.
by Kristof77
Compare to how the poor live today to how the wealthy lived 80 years ago, or how royality lived 250 years ago. I’d argue the standard of living for everyone has been improved and I see no reason why that trend wont continue. Of course those with the means will be the first to benefit from the lastest and greatest but eventlually that all trickles down. Id also argue that trickle down effect (which has nothing to do with trickle down economics) is also increasing in pace.
by Dr. Bloggs
I see poor people in the US with iPhone 5s all the time…
by Cybernettr
…And in India there are more smartphones with Internet access than commodes. Personally, if I had to choose one or the other I would choose the latter.
by JC
I think enough people are finding strong altruistic drives within themselves and that technology will enable them to help the world’s extreme poor to a degree never seen before in history. One of the great stories of our times, written in future history books, will be the numbers and percentage of such poor who are raised out of that abysmal level of poverty.
by Vin
Cutting edge is a two-edged sword: can still be dangerous and experimental. Long term ramifications can only be estimated to some degree? Maybe many rich people shall be dying and suffering before some common ideal across humanity is reached.
by Christopher Feyrer
I will feel better about Ray’s “so inevitable it bores him to repeat himself so much” future when I’m assured compassion, not blind Darwinism, will guide the future. Ray’s refusal to confront and dismiss these arguments concerns me. Or perhaps the dismissiveness instead reflects all of our eventual irrelevance?
by Jack Richardson
It is not only 3d designs that will maintain ‘accelerating returns’. There is a transition to ever greater optical designs, data readouts from the spin of the atomic nucleus, and advanced software that uses quantum dot computing. The research already in place can support another four or five doublings.
by Khannea Suntzu
This sounds like fun! … however…
Ray, what are the consequences for employment in the current (rather stubborn, reactionary, militant, short-term thinking obsessed) corporatized world of Moore’s Law?
Why does “Libertarian Transhumanism” completely ignore (or belittle) the massive human tragedy that automation and robotization directly triggers?
Do you advocate policy/political solutions (other than blindly trust on the ridiculous and long-falsified “trickle down” nonsense) to repair/compensate the massive societal disparity that is already emerging, let alone when billions of jobs will become irreversibly replaced by automated machines?
..how do you retrain tens of millions of service industry / retail / fast food workers in the span of a few years? What else does society have to offer these people than Ghettoes and Dole Yeast?
…you will probably say something palliative at this stage that “we” will merge with machines” ? Is this “we” also applicable to the 99%, Ray? Or is it Cake for everyone else?
by Mark
EM Foster wrote about the difference between a preacher and a prophet. A preacher will say how he wants things to be, a prophet will say how they will be, like it or not, for better or for worse.
Obviously I don’t know Ray’s thoughts on this, but having read some of his work, I think he’s predicting how things might be, whether we like it or not. Could the internet and technology be the end of us? Yes, it could.
by pauldenice
To the question you ask: “Why does “Libertarian Transhumanism” completely ignore (or belittle) the massive human tragedy that automation and robotization directly triggers?”
I have tried for many years to attract the attention of “social NGOs” to this risk, however with no success: the reaons being that most such organisation members have no interest nor knowledge on technology evolution and impact on their own constituencies. Most have a “law” background or went to humanities during their university years. As a result most of them can’t understand or worst,refuse to follow sciences and technologies evolutions, not even through reading the most basic popular science and technology magazines. such as “Scientific American in the USA or equivalent magazines elswhere in the world.
by Paul Marsh
Very well said!! Thanks for raising the issues and asking the questions!
Martin Ford in his book ‘Lights in the Tunnel’ said that although Artificial General Intelligence is and will be a great benefit for mankind, it will also be economically disruptive. Eric Drexler in his book ‘Radical Abundance’ suggests that although nanotechnology is and will be a great benefit for mankind, it will also be economically disruptive. AGI and nanotechnology will cause massive unemployment and will tend to add to the already-at-a-60-year-high income inequality in the U.S. The U.S. and the world need to start thinking and planning for how to deal with these rapidly approaching economic problems.
by Michael E. Arth
Robots and automation will quickly render most jobs obsolete. At some point, we should have a guaranteed minimum income (GMI) and national health care. Look at it as a royalty on the common technological heritage that is no different than distributing the benefits from any other commonly-shared heritage (like oil found under public land).
The GMI should also include free education, because smart (and enhanced) people will be able to give something back to society. We will also need choice-based, marketable birth credits to stop or reverse population growth because people will live longer–at least until they upload their minds into cyberspace or whatever it is that we will do. A GMI will help smooth the passage from human to post-human.
by karl
I think you may need to caveat your statement.
Robots and automation may render most manual labor and manufacturing jobs obsolete. The service industry , entertainment industry , and the professions will likely remain unaffected until strong AI
by Jack Richardson
After three more doublings, there is likely to be significant changes even in the service, entertainment, and professional worlds.
One can imagine going to a McDonalds, placing one’s order with a realistic human on a screen, having it cooked and delivered automatically, and paid with a pass of one’s smart phone.
In the entertainment world, it will be almost impossible to tell the difference between computer-generated characters and real people.
In the professional worlds, legal research as well as many other kinds of research will be done by intelligent apps. In hospitals, more and more operations will be done in a robotic environment that will be more precise and dependable. One could go on and on.
by axel
Most jobs that existed 70 years ago are rendered obsolete due to new technologies. New technologies mean new jobs. Take for instance augmented reality, that is on the verge of a break-through, it will open up a whole new world. At least for developers and marketing. :)
by melajara
And, ultimately, eleven dimensional computing, i.e. computing through the full multiverse!
see http://www.youtube.com/watch?v=jI50HN0Kshg
by Bri
Mischio is really cool. I don’t quite understand how eleven dimensions can be utilized by us as a computer.
by Boy
Three dimensional computing, i hope it will be a reality …
by fellowhuman
It already is. Flash memory in 3d design (3d NAND) already exist and will be available to customers in 2014. Stacked CPU layers are also in the pipeline. Developed by 3M and IBM.
by karl
Thru Silicon Via — google the terminology
Basically stacking chipset components vertically and connections are through viaducts in the stacked components.
Cuts power consumption and heat at the same time (less voltage needed because distance between components is negligable — and with less voltage comes less radiant heat)