Slashdot Mirror
Michio Kaku's Dark Prediction For the End of Moore's Law
nightcats writes "An excerpt from Michio Kaku's new book appears at salon.com, in which he sees a dark economic future within the next 20 years as Moore's law is brought to an end when single-atom transistors give way to quantum states. Kaku predicts: 'Since chips are placed in a wide variety of products, this could have disastrous effects on the entire economy. As entire industries grind to a halt, millions could lose their jobs, and the economy could be thrown into turmoil.'" Exactly the way the collapse of the vacuum tube industry killed the economy, I hope.
Noone will take a disaster prophecy seriously if you can't even be bothered to pair it with some planetary alignment or ancient calendar.
The major difference being the tube/valve industry was done in by the transistor - i.e. we had a viable replacement that was better. The problem with the transistor is that we don't (yet) have a viable replacement.
Take one paragraph, take it entirely out of context, and add a sarcastic line, and miss the point entirely. Timothy didn't even bother to RTFA perhaps? Kaku has a pretty high degree of credibility, he is an excellent write. Shame on you, go and read what the man actually says.
Every time a new technology comes along, there is always this lingering fear that it might kill the economy, but as it turns out every time, new jobs are created and std of living of ppl is raised! so chilllax!
I predict a dark future for Michio Kaku's new book.... namely, the bargain bin.
Take Nobody's Word For It.
Software developers are going to have to consider increasing efficiency as they make their wares more complex! And we might have to actually implement concurrency research which is under two decades old!
Who knows, we might even end up with the responsiveness of my RISC OS 2 Acorn A3000 in 1990.
Does anyone pay any attention to Michio Kaku? He isn't quite as much a woo merchant as Deepak Chopra, perhaps one could compare him to the likes of Henry Stapp or Fritjof Capra.
Apparently people can't:
make cluster computers
make boards with multiple chip sockets
make extension boards that go in PCI or potential future slots
use distributed computing
[insert many other ways to get around limited processing power]
Man, we sure are screwed in 20 years time, computers will start rebelling against us because we can't make them smaller than the boundaries of the universe!
On a more serious note, this is retarded. Period.
20 years is a VERY long time.
By then, we'd probably actually have the beginnings of working quantum computers that are useful.
By then, we'd have almost certainly found out how to get around or deal with these problems, possibly even taking advantage of quantum effects to reduce circuit complexity and power needs.
Who knows, but i know one thing for sure, the world won't end, life will go on and usual, and this book will still be shit.
Michio Kaku is not necessarily the best in his field, mediocre at best, but he has the biggest voice. I was talking to an older woman awhile back and she is a devoted fan of his. I asked her what she knew of him other than that he does "layman's" break down commentaries of Physics for the discovery channel and she actually thought badly of me for trying to undermine her opinion of "the top physicist in the world today". Well, that's definitely HER opinion and not mine. Just because he has a big mouth (media wise) does not make him remotely right on anything is the point I'm trying to make here. oh, I just got it- Now I understand Politics lol
"Computers are a lot like Air Conditioners" "They both work great until you start opening Windows"
Sooner or later it will come to an end, but it will come slowly as the challenges rise, the costs skyrocket and the benfits are lower due to higher leakages and lifetime issues. And designs will continue to be improved, if you're no longer constantly redesigning it for a tick-tock every two years you can add more dedicated circuits to do special things. Just for example look at the dedicated high def video decoding/encoding/transcoding solutions that have popped up. In many ways it already has stopped in that single-core performance hasn't improved much for a while, it's been all multicore and multithreading of software. Besides, there's plenty other computer-ish inventions to do like laying fiber networks everywhere, mobile devices, display technlogy - the world will still be in significant change 20 years from now. Just perhaps not on the fundamental CPU code / GPU shader level.
Live today, because you never know what tomorrow brings
Yeah, maybe we should stop the waste, and employ human operators to send telegraphs like they did in the good old days, scribes to write documents by hand....
Don't quote me on this.
His view is based upon the chip and not on the device.
What I'm seeing is folks (manager types ) using their iPhone as their business computer - eliminating the laptop and even their desktop. They're on the move and what they need is a portable communications device that also has some other apps.
Spreadsheets? That's for the back office staff. The same goes for anything else that you still need a desktop/laptop for.
So what's my point - desktops and laptops are becoming a commodity back office device (like the typewriter in the past) and the demand has stabilized and as far as business apps are concerned, there isn't any need for more power - bloatware aside.
To head off the "64K of RAM is all anyone really needs" comments, that was then, this is now. Back then, we were at the birth of the PC revolution. Now, we're in the mature commodity stage. Will we need more power in the future? Yes. But at Moore's law increases? Nope.
The future is efficiency, portability and communication.
PC's are inefficient for most uses; therefore, there won't be any "death" or "economic" destruction - just some "old" companies hitting the skids (Dell) or vastly changing their business if they can (HP).
Really, IT has far more wide ranging applications than a fridge and can create new ways of doing things, these may not always be better but a good proportion of it is. People who think that IT is a waste are usually the same people that think the space program is a waste or that education is a waste. Progress has to come from somewhere, it is not magiclly pooed from the buts of celebrities or political figures as they dance about appealing to the masses.
I take it that you are too young to pay the electricity bill... Basement? Cooler down there?
Deleted
A "new refrigerator" is, supposedly, more efficient than the last one. The emergence of IT made entire armies of secretaries, messengers, archive managers, human computers etc obsolete, changing society profoundly. The comparison to an iterative development of an existing technology strikes me as moot.
If we actually do need all this extra processing power, we can always build a 1x1x1 km big cube of atom sized transistors, and then time share it with the planet.
Just for example look at the dedicated high def video decoding/encoding/transcoding solutions that have popped up. In many ways it already has stopped in that single-core performance hasn't improved much for a while, it's been all multicore and multithreading of software. epayroll
Atoms are composed of protons and neutrons which in turn are composed of hadrons and quarks. So we can used hadrons and quark based chips so Moore's law can go on.
There has always been a suffering factor built into changes in technology spilling over and causing changes in society. Usually the suffering has been rather confined. The buggy whip workers were not such a large group of workers that the new automobile market destroyed. But now things are different and less predictable. A great example is in office staff eliminated by the cell phone. As cell phones took over the small company was able to get rid of millions of girl Friday types that had answered the phones and greeted walk in customers in the past. We are now displacing workers so quickly that economic chaos is descending upon us. After all, earners pay taxes whereas the unemployed consume tax dollars.
Back at home the US has no clue with what to do with labor. Just as we saw people suffer loss of life and health trying to save victims of 9/11 now we see brave,skilled, laborers in Japan entering death zones in order to save the public. I am certain that all that really matters to the republicans is that those workers do not have a union nor want help while they die from radiation sickness just as they will not take care of firemen and police with busted lungs and cancers from the WTC rescue efforts. Frankly if we suffer another attack perhaps our public employees will smarten up and stay far back from such catastrophic events.
Imagine if someone else came up with a "new refrigerator" and the efforts on maintaining the "new refrigerator" came to suck up 10% of the economy.
How big of an LCD will this fridge have? Will it have USB 3, Thunderbolt or Gigabit Ethernet? How about WiFi, a full Bluetooth implementation or this new fangled NFC stuff? Will my better half be able to hook up a scale that not only weighs me before I open the fridge but after to see exactly what I took out? Will a pre-recorded movie play that tells me I shouldn't be eating whatever I just took out, reminding me of my diet or just asking "are you going to bring me one, too?" What about commercials? "I see you're running low on Pepsi 3000! You should go buy some more Pepsi 3000! Now!! But wait, it's a long trip to the store and you may get thirsty - why not have a Pepsi 3000?" Will I then here her voice telling me "put that back! Have some fruit instead!"?
The end is never really the end. Just the beginning of the next thing. What makes anyone think that the quantum level is the end of smallness?
Admittedly it sucked but they did have a class in parallel programming that had some of the guys in the physics dept helping out. They did point out that eventually coders were going to have to do parallel programming and break up computation into chunks that could calculated simultaneously because of a little thing that would stop any single processor from getting faster. That being the speed of light.(IE you can only move data around so fast and when you have to move data around faster than light you have a big problem.) Admittedly this was over a decade ago and come to think of it most modern cpus are multi-core so looks like he had a point.
Even if Moore's law come to an end, we can still improve the performance of the systems via parallelism.
He made similar economic predictions in the BBC Horizon episode "The dark secret of Hendrik Schoen" (2004).
That was the day I lost all respect for Kaku. His economic predictions are moronic (there will always be change, abrupt changes in what creates wealth), and in that Horizon documentary his comments seemed ludicrously off track as well.
A really good tip an AC gave me ages ago: change troll and flamebait mods to +5 and browse at -1. A lot of good comments get modded off for people disliking them. 0 isn't the end of all quality, you know. Also: holy crap, Slashdot's over 2400000 accounts? I suddenly feel ancient, and I'm only 22!
The role of the writer is not to say what we can all say, but what we are unable to say. -Anais Nin
Once hardware stops growing exponentially, software design and compilers will be optimized for performance. Little regard was spend in the last years to that because hardware would solve the problem. Once hardware performance stops to grow exponentially, software design and infrastructure will grow exponentially -- there is room there for at least another 20 years. Around that time quantum computing will change the world and will give rise again for yet another innovation cycle.
Even with transistors the same size, there are so many avenues to explore in processor design. Just off the top of my head, how about a memristor-based analog co-processor for tasks like facial detection or language/speech recognition. How about processors with asynchronous clocks, or clockless designs. Sure, they're harder to build, but once transistor sizes fixate, might as well spend the effort because designs will have a much longer lifecycle.
I don't think that was his point either, although "waste" is a bad wording. Consider this: back in the days everyone were farmers. Farming was, and still is, a requirement for society to go around, but in modern developed countries we don't need to allocate all our resources into agriculture to keep it up. If we can keep IT in the future with human resource consumption at a minimum, it's a win-win situation for everyone. Disregard GDP and all other fancy abstraction; humans are, after all, society's only resource, and as of such it's a highly precious one.
Wait, does that mean I've been wasting the 20-30% of my budget that I spend on food? I sure am going to miss it. Oh well, at least my pastime of throwing dollar coins at drains only costs me about 2% of my income and is therefore not wasteful.
You know, there is a difference between trolling and pointing out the flaws in your reasoning. Just saying.
This guy is trying to establish himself as some kind of authority on futurism, but I just perceive him as an attention whore who actually contributes very little. Maybe I'm alone in thinking this, but his TV series "Physics of the Impossible" was one big self-aggrandizing marketing gig. I barely made it through two episodes that essentially consisted of the professor rehashing old science fiction concepts and passing them off as his own inventions. Every episode ended with a big "presentation" in front of dozens of fawning admirers. Before the credits rolled, they made sure to cram in as many people as possible saying how great and ground-breaking his ideas were. It was disgusting.
Are there physical limits to Moore's law? Sure. We already knew that. Circuits can't keep getting smaller and smaller indefinitely, and we have already run into the limit on reasonable clock speeds several years ago. And despite this, the computer industry hasn't cataclysmically imploded.
Nut case.
While parts of technology might stop progressing as fast, other parts of technology will start getting optimized, to get over the halting of that other part. So if hardware stops getting faster, people will start optimizing software (which is currently extremely inefficient), until we get a better HW/SW tech at some point later in the future. There's a very nice comment on the Amazon page of the book by JPS, give it a read.
I fail to see how the inability to make smaller computers for ever will stop us from still needing them by the boatload.
From weird analogies and a certain amount of misunderstanding things the excerpt draws strange conclusions.
a) Misunderstanding how the frequency spacing relates to required number of cycles: The correct assumption would be that if light has 10^14Hz and you restrict yourself to single-octave circuit (for the sake of simplicity: lets say 10% relative bw circuit), then you can if you "cram" ideall of modulate fast enough, 10^13bits*log2(S/N) bits per second. so probably 10^14bits/second - that is a lot.
b) limits to Moores Law: Moores law is an economic law. There is no physical limit which i see which can be reached technologically until 2020 (in mass production). There is a technological limit to what can be produced, but going in the third dimension and new materials will give opportunity to continue on the same course for a while. If you look at what physicists are currently looking at, you realize that the end of silicon/metal/oxide technology will not be the end of Moores Law or classical computing
c) "on the atomic level i cant know where the electron is". As it happens to be i work on quantum computation and i really hate to explain that: If you arrange a specific situation, then you cant know where the electron is on the atomic scale. If the statement would be as general as he makes it, it would be impossible to have different chemical configurations of the same stoichiometric mixtures. SIngle-molecule electronic/magnetic configurations. The quantum tunnel coupling in single molecule magnets between states can be designed, and i dont see a specific reason why it should be impossible to realize single molecule devices in which tunneling does not play a role
d) he does not understand FETs AFAIU
e) contrary to his opinion, very thin 2DEGs exist and i dont see a reason why upon (finding and) choosing the right layers, the confinement can be very steep in the third direction (not infinity, but also not requiring more than 50nm thickness)
The funny thing is that he forgot what already is and probably will (there *may* be ways out, like superconductors or ballistic transport but don't bet on it) really be a problem for all classical/room temperature computers: heat. While the designing smaller elements may be possible when using the right physics/technology, reducing the capacitances of lines (associated with an energy loss in the line resistance per switching) will be difficult. Once we *really* stack in the third dimension it will need a lot of clever computer scientists (and maybe mathematicians) to reduce thee needed interconnects, since otherwise stacking the third dimension wont give us anything besides memory capacity.
So to conclude: i believe until 2050 the definition of Moores law will be obsolete. but it will not break down because we are unable to make circuits smaller, but because it may be too expensive to make them smaller or powering and cooling the circuits may become impractical. We probably will have a replacement of moores law by an equivalent scaling law for power per switching.
Another pseudo-economist out to tell us that an increase in productivity and a lowering living costs will be a net loss for society. Michio Kaku can you please take an economics 101 class before writing a book about the economic impact of anything. The general population is already economically illiterate and this only fuels the problem. Thanks.
The transistor will be replaced by 3d self organizing molecular circuits. Watch this ted talk by Ray Kurzweil from 2005, where he explains the whole paradime shift issue. http://www.ted.com/talks/ray_kurzweil_on_how_technology_will_transform_us.html "Inventor, entrepreneur and visionary Ray Kurzweil explains in abundant, grounded detail why, by the 2020s, we will have reverse-engineered the human brain and nanobots will be operating your consciousness." - Thats right, nanobots will be operating you consciousness.
DAD?
Damn ambulance chaser. I love physics; I despise Michio Kaku. He gives physicists a bad name.
He would have done better if he would have gotten his Phd in home economics.
Kaku is an embarrassment. In the mid/late 90s he presented himself as a "nuclear physicist" to the major news outlets (he is no such thing-he's a field theorist) and jumped on the till-then fringe protest movement opposing the launch of the Cassini mission. The opposition was based on the idea that the nuclear batteries on the probe posed a danger in the event of a launch accident. Nevermind that there had previously been launch accidents with the same battery type (military sats) and the ceramic uranium cores were simply recovered and _reused_ because they're practically indestructible. (The batteries are just warm bricks. Low level uranium fission keeps them cooking and thermoelectrics generate the juice. There are no controls to go wrong, no parts to break, nada. That's why they're used. The ceramic itself is terrifically tough.)
Anyway, Kaku saw the cameras and the bright lights and decided that he was a nuclear physicist and start spouting all sorts of total nonsense to frighten the unwashed masses. He has a long history of pretending to know things. Google "Kaku evolution blather" for another example. I watched him give a seminar once while I was in grad school and I just spent the hour squirming in embarrassment for him and his self-aggrandizement.
Yes, I loath the man. I'm a physicist and he just perpetuates the image of people in my field as hubristic egoists. He needs to be shouted down and run out of the media. There are lots of really good popularizers out there (DeGrasse-Tyson, Greene, etc) who, yes, need to establish a public presence to make a career, but who are also actually interested in facts and science and education and know their own limits.
Just read the article, haven't read the comments yet.
Moore's law as far as CPUs and GPUs has already slowed down considerably this entire decade. As far as memory, so far as memory the chips aren't that thin yet. What this means is what everyone has been saying for a long time: more cores, more ram. More cores means applications need to be parallelizable. That's at least a one time overhaul of most of the world's code base.
Lets assume hardware improvements in general slow down. This leads to a hardware situation closer to what we had in the 1980s:
a) Because hardware is stable operating systems and applications can be written more efficiently to take greater advantage of the hardware. That means refactoring high level code into lower level code to get speed ups. The popularity of Java is basically based on rapidly changing underlying platforms, make platforms stable and we have language revolution with much less hardware abstraction. Compilers will get faster as well.
b) Because hardware is stable computers don't seem like as much of a disposable item. Getting a good quality system to keep for many years makes economic sense. So we can get a one time boost as people move back from $700 computers to $3000 computers, where they expect to get 10 years+ out of their computer.
c) In the area of cell phones we could see the same thing. While cell phones are too breakable to ever become extremely expensive (though there are people who get expensive cell phones now: https://store.vertu.com/en/) if the platform were to stabilize we could see much richer client applications. If you expect to be on the same cell phone (with just hardware replacements) for a decade your willingness to buy expensive software goes up.
So lets say I don't agree with 2020, because around 2020 is when you start to see everyone upgrading. Which of course leads to software with much higher system requirements which drives more upgrades.... But maybe 2040 we have a stagnant computer technology industry if nothing interesting happens. I guess that could happen but I dont think its likel. However even if it did, this creates another advantage. You now have stagnant hardware, stagnant operating systems, stagnant languages, stagnant applications. An environment where corporate computing and custom code becomes a great value. And that is a huge and ever growing code base. So now we are out around 2060, where the industry is in maintenance mode. So the question is: between 2010 and 2060 do you think no one is going to come up with a really good idea?
Moore's law is not so bad compared to the bigger picture of running out of resources. An economy that assumes exponential growth will be thrown into turmoil, and millions^Wbillions will lose their jobs.
Escher was the first MC and Giger invented the HR department.
Here is a news flash. I have it on good authority that
- eventually Moore's law will fail, and
- the world will continue to roll through the void. Life will go on, and we will not burn our Mac Book Pro's for heat, nor turn our rack-mounted servers into crude dwellings.
Huh? Because computers stop improving exponentially doesn't mean that there won't be continued, and additional, use for those we have. And there are lots of other avenues for innovation for storage, input, output, etc. that might not result in the raw performance gains we've become accustomed to - but will still be innovative and drive research and sales.
I'm not a genius, but, since we don't know all the laws of the universe, how can we possibly make such a prediction? 20 years is a long time, a lot of things will be discovered. But we still won't have flying cars tough...
Lo and behold, for I am a sig!
Worth noting this table? Specifically the overall rows at the top for men and women. Income for men has been flat since 1970 when adjusted for inflation. All the income gains have come from women entering the workforce, going from partial to full employment, and/or the gradual elimination of sex discrimination which drives down wages. One could also argue the cost of living has actually risen faster than official inflation measures, especially when one includes the additional costs necessitated by both partners working full time. (Day care, outsourcing tasks like cleaning and yard work, etc.)
MOS transistors were developed prior to 1970, but not by much, and they didn't really start catching on until the 1970s. Now I'm certainly not arguing causation here, but by the same token I'm not sure it's valid to suggest (via sarcasm) that the move from vacuum tubes to transistors ushered in a new golden era of prosperity.
Humans as the most important resource is the old economy ... due to overpopulation and automation the most important resources for a society are slowly becoming natural resources. For the competition between first world countries it's already true, the country with the highest median wealth has a trade surplus based on oil. The comparative advantage of an educated labour force is diminishing very fast. So no, all those people freed up in IT will not flow towards creating more wealth in other ways. It will just lower the wages of Wallmart greeters.
Thanks alot SALON, your advertisement caused Firefox to cause Windows to bluescreen [ BBCode 10000050, or a "page fault" in memory ]
I don't know. I'm fairly certain that she was alive when we started...
Actually, you brought up a problem the *excerpt* doesn't even get into - the whole cores & threading tussle. (But this is from a whole book, so we can't speak to the whole contents!). It might mean that a 64 core computer might only use some 4 cores because every dev can't always work in the complexity of parceling out tasks to an undefined number of cores and have it optimize every time. In that sense we might lose ground against Moore's law early.
Maybe it would take a hardware plateau for the big software corps to finally knuckle down and do a ruthless rewrite for optimization on existing features, and get a software side quad increase in effects.
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine
I can't remember the book I read it it, but the author made the claim that if you look at the part of Moore's law about human computing power, i.e. doubling our ability to compute numbers every 18 months, and temporarily ignore other parts of Moore's law it holds true going back to pre-history times.
In other words silicon computer chips took up where the abacus and solid state circuitry left off, etc, going all the way back to putting marks in the sand on the ground.
If we get to the theoretical limits of silicon something else is going to take its place. That's not a matter of faith because history shows it always happens.
He forgets the human element. It does not matter if Moore's Law collapses, we will continue to upgrade and buy simply because "they" tell us this years model is better than last years. I offer as an example the iPhone and the latest release. It had been reviewed, tested, dissected and obvious problems were pointed out such as the antenna problem, etc. Did people wait to see if it would be corrected? No, they rushed out and bought them like there was no tomorrow. We will throw away last years device and buy this years device simply because "they" tell us we need to.
Kaku's predictions will have some effect but not that much.
What's worse?
Free-as-in-beer Energy. The Sun is radiating for free vast amounts of energy. All we need is a highly efficient way of capturing this, converting it into electricity, and storing it. Once that is done, human labour will become prohibitively expensive compared to machine labour. Aside from the Sun there's fusion energy - once it's up and running "just add water" and the output energy is free.
Singularity. Enough said, once the first AI wakes up, more will follow.
Doomsday. We're still not out of the danger zone (see Doomsday Clock) for a large number of reasons; chances are low but it can't be ruled out.
He might as well start predicting the weather one year ahead.
I mean I agree Kaku is a blight on science. (I mean his explanation of why E=MC^2 on the science channel basically amounted to "That's what Stone Cold Al Einstein said so.") Still Tyson? I mean we are talking about a guy who romanticizes so much how black holes suck everything down around them. Honestly, the way he puts it he makes it seem as though if a black hole were to go through the solar system getting sucked in would be the major concern. (You know, not mentioning that you'd have to get fairly close for that extreme gravity effect to take place at all. The real concern is it completely messing up the orbit of earth and causing us to freeze or fry. Of course he doesn't say that because it doesn't sell books. Oh well, at least they weren't as bad as when I saw Kip Thorne on one of those channels explaining quantum entanglement. (He was technically correct and yet he would have totally confused any lay person who didn't know anything about it. He made it sound like the screwed up version of quantum entanglement.)
Did you know 80 to 90% of the moderators on slashdot wouldn't recognize a troll even if one dragged them under a bridge.
The world now functions under an economic system based in the principles of Expansionism. Whether it's tech, banking, real estate, or business in general, all are predicated on an expanding market. Unfortunately, the world is a finite place and this entire model is destined to eventually crash catastrophically, likely taking the society along with it. Sadly, there is little stomach among either the leaders or those led to even acknowledge the problem, let alone contend with it.
There are parts of atoms that already lend themselves well to binary behavior. Of course, figuring out how to manipulate them would be a challenge. But of course, Moore's law pertains to number of transistors that can economically be put on an IC right? Well, if you had an infinitely big chip, you could put an infinite number of transistors on it, so there is no theoretical limit. It just is a question of economics. Additionally, I don't think we have done much playing in 3D as far as chips go. We have mostly been putting transistors on a flat plane. There is another whole dimension to play with. For example, the latest Sandy Bridge 2600K has nearly 1 billion transistors. For simplicity's sake, assume they are laid out 30,000 X 30,000. Now, what if another dimension were added. Multiply by 30,000 and all of those transistors are still about the same distance away from each other. That is 27 trillion transistors.
Just figuring out how to cool such a chip will keep the economy afloat for a decade or so.
If you are not allowed to question your government then the government has answered your question.
Sheldon Cooper: [Talking about Penny's proposed home-based business] If you took advantage of modern marketing techniques, and you optimized your manufacturing process, you might be able to make this a viable business.
Penny: And you know about that stuff?
Sheldon Cooper: [patronizing] Penny - I'm a physicist. I have a working knowledge of the entire universe and everything it contains.
Penny: Who's Radiohead?
Sheldon Cooper: [with facial tic] I have a working knowledge of the _important_ things in the universe.
I predict that when Moore's law "slows down" over a decade or so, the industrial and intellectual power can move to software and improve algorithms and system efficiency (for a while).
We see the same now with peak-oil on the horizon that the internal combustion engine gets more and more optimization and is driving car sales of all manufacturers. Something that had been a waste of time and money only 12 years ago (see VW Lupo 3L and Audi A2).
The number of people claiming the end of Moore's law doubles each year.
Moore's Law doesn't say chips get smaller, it says they get 2x the transistors every 18 months. It doesn't matter if chips hit a wall where transistors can't get smaller. As long as they continue to get cheaper, we will simply start growing by making bigger chips, or more chips, or 3D chips.
Further, we can get speed other ways: no moving parts, better software design and optimization, simplification. iOS v4.3 on a single core 1GHz ARM feels faster than Mac OS v10.6 on dual core 2GHz Intel because of factors other than Moore's Law.
And we get economic growth in other ways. Just getting out of the Wintel monopoly is better for business than Moore's Law. Just moving from IE6/FlashPlayer to HTML5 is better for business than Moore's Law. The Web jumping from PC-only onto smartphones in 2007 is better for business than Moore's Law, and had more to do with software than chips, which is why Apple did it, not Intel. These are all well underway, but many benefits are yet to be realized. The computer business has mostly lacked real competition for decades. It's getting healthier now and less reliant on Moore's Law to make things better for everyone every 18 months.
So in short, even if we hit a wall in Moore's Law in 2030, we won't care like we would if we had hit it in 2000.
A "new refrigerator" is, supposedly, more efficient than the last one. The emergence of IT made entire armies of secretaries, messengers, archive managers, human computers etc obsolete, changing society profoundly. The comparison to an iterative development of an existing technology strikes me as moot.
A very interesting expanded comparison between IT and refrigerators can be made. The introduction of refrigerators also changed society profoundly (though perhaps not quite as profoundly as IT). The ability to ship food long distances and store it for long periods of time throughout the supply chain, right down into the kitchen, the invention of new food products (Birdeye's flash frozen vegetables, etc.) had major economic and social implications. But this was in the first half of the 20th century, long before the life experience of /. readers.
And since 1972 there has also been a genuine revolution in refrigerator technology. Prices of refrigerators have plummeted, and efficiency has sky-rocketed. This chart only takes us to 1997, but it shows a near-tripling in energy efficiency over 25 years, and the progress has not stopped since then. A new 18 cubic foot refrigerator uses 350 kWh per year, which is an average energy consumption of only 40 watts. Most PCs use more power than this, even taking into account the long periods of idling.
I suggest that computer processor energy consumption needs to follow a curve like that of refrigerators as a share of the national energy consumption - reverse its still climbing share of national energy consumption, and begin a long decline.
Second class citizen of the New Gilded Age
I am so sick of seeing Michio Kaku all over the place...
It made sense back when he was talking about string theory. He's a physicist, after all. But these days he's just some generic scientist who's more than happy to show up on TV and talk about anything even vaguely scientific.
Did you see him commenting on the whirlpool formed after the earthquakes in Japan? Because a physicist is obviously the most qualified person they could find to talk about ocean currents and plate tectonics and whatnot.
What makes Michio Kaku any more qualified to talk about Moor's Law than I am? It isn't like he actually knows anything about microchip fabrication or economics or industrial processes... The guy is a physicist.
"Work is the curse of the drinking classes." -Oscar Wilde
Or the cost of biofuels... Humans contain enough lipids that converting them into fuel shouldn't be too difficult.
Like others I believe Kaku is wrong. Here is my prediction:
Within the next 20 years massively parallel processing will become more and more common, machines with a few dozens, hundreds or even thousands cores will be the rule, and programming languages / compilers will be able to automatically turn sequential programs into parallel ones whenever this is possible. Almost all practical computing problems and needs will turn out to be highly parallelizable. The impact of this change on economy will be zero. Computers will never stop to become faster and faster.
In 50 years from now or earlier our massively parallel conventional machines will be substituted by quantum computers. These will first be available to governments and big companies and within a short period of time will be miniaturized and become available and affordable to end consumers.
what Fox news is to journalism.
That means that we'll finally be working on improving the software for improvements over just writing code and letting the faster hardware overcome any performance issues.
Bryan
Society is not going to run down because we need smaller electronics. It's nice to have portable devices with enough compute power to play grandmaster-level chess, but most people max out running Angry Birds and watching TV.
Society is going to run down because we need energy sources. Every energy source we have today is a minor improvement on technology at least 50 years old. The non-renewables are running out, fusion didn't work, fission isn't looking too good right now, many of the really good wind sites already have wind farms, solar doesn't work at night, and growing crops for fuel competes with growing food. The future will run at a much lower energy level than today.
That idiot? Michio Kaku has made an entire media career out of over-simplifying science, and ignoring any and all complexities that arise in the course of doing real science. The man is a loon, and anyone taking anything he says about the future of technology seriously is in dire need of psychiatric attention. Put him in a room with Ray Kurzweil, and watch the looniness explode.
Anyone else read doom and gloom industry and business prognostications with unbridled glee? I fail to see how having technology become so ubiquitous that everyone can afford it could be a bad thing. Yes for technology companies and hardware manufacturers its going to mean a shift from boutique to commodity but for society in general it will be a great thing. Ideally we will eventually reach a time when basic needs of sustenance, communication and shelter are just there and when that happens most of our "jobs and businesses" will likely be long since rendered useless to society. People will be free to pursue their own interests and even with the "lazy" that some now will complain about society will be better for it. Enlightenment and advancement will one day trump wealth and power, despite how many fight to keep it from happening.
Wrong! Does not matter...
http://www.ted.com/talks/bill_joy_muses_on_what_s_next.html
Technologist and futurist Bill Joy talks about several big worries for humanity -- and several big hopes in the fields of health, education and future tech.
amusingly, that only confirms Kaku's prediction.
If your existing refrigerator is perfectly good, then what incentive do you have to buy the NEW refrigerator?
People periodically buy things... appliances included... simply because they want new. More shiny, more function, or in the case of wives, simply a matter of "That look is so ten years ago". In many cases, people buy new things for the sheer pleasure of doing so.
Life is hard, and the world is cruel
Kaku is a bullshit TV presenter, no science here
Now he's an economist? He's a pretty crappy, self-promoting physicist. I suspect he will be a crappy, self-promoting economist.
Wait i just got a glimpse of the future!
Michiu kaku will predict that one day we wont collectively need or wany the quantum flux statesnproduced by michiu kakus continued existence so well wish him out of history. George boori, art bell, richard hoagland, and bill nye the science guy will surround kaku and atempt to sbvert his quantum presence with the ghosts of half of albert einstein and half of mister wizard. They will be unsucessful so we will chamge quantum states to avoid paradox ehich is impossible, and just wish him into an iceberg made of junk, science, and washed-up fringes. Michiu kaku will be right but dead, the resultung revelation inspired paradoxialironic tension will result in a mass conclusion that michiu kaku was wrong but died anyway.
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
Their balance depends on culture and technology:
* Subsistence
* Gift
* Planned
* Exchange
* Theft/Parasitism
So, as Moore's Law comes to an end, having moved up and S-curve, we could see a shift between these types of economies. The exact balance somewhere would depend on the culture.
See also Marshall Brain's on-line book "Manna" for some ideas of, say, a basic income might look like.
Related:
http://peswiki.com/index.php/OS:Economic_Transformation
http://www.marshallbrain.com/manna1.htm
A 21st century issue: the irony of technologies of abundance in the hands of those still thinking in terms of scarcity.
Garbage. Once Moore’s ‘law’ (it is not a law, only a prediction meant to hold for a while only) stops holding, we can just go for more efficient software and architecture. Faster hardware has enabled quite some lousy software and hardware architectures around. We are seeing some of this already, in ARM and GNU/Linux instead of MS Windows and Intel; stuff such as Lisp machines, for instance, could given even bigger gains.
Leandro Guimarães Faria Corcete DUTRA
DA, DBA, SysAdmin, Data Modeller
GNU Project, Debian GNU/Lin
Yeah, but at least 40% of those are Micheal Kristopeit.
I thought Singularity would come earlier :D
Everybody seems to have forgotten this one in particular, in which the Cassini-Huygens mission held the seeds of mankind's destruction.
Based on his track record it simply does not make sense to give Kaku access to any form of publicity. He may have done some good physics work at one point, but he left rationality behind when he tried to stop the Cassini mission with wild-eyed hand waving.
BTW, nothing in the above post should be construed as questioning Dr. Kaku's medical history, academic qualifications, professional competence, or personal behavior, only his motivations and past activity as a highly visible public figure in scientific and technological fields.
We already have computers and storage systems that exceed the capability of a human.
The reason that we do not seem to have any AI more powerful than an ant is software, not hardware.
Moores law is great, but the problem we seem to have is that our ability to construct complex software does not seem to mirror the same law.
There's one main flaw in this argument. If we can no longer double the number of transistors in a given area every 18 months, does it really matter? Does the world fall apart? I think not.
Existing chips work. They won't stop working just because we can't make smaller/faster ones. Yes, we are accustomed to seeing newer/faster/smaller chips on a regular basis, but it isn't necessary for the world to continue. Even if we could never make a chip more complicated than the ones we have today, we would be fine.
Massive over-reaction / sensationalism.
"Tea. Earl Grey. Hot."
Today everything becomes obsolete so fast, nobody wants to make a long term investment in a high tech device. But if we do reach a limit, and the advance of electronics grinds to a halt I think a whole new economic model will emerge. Where people will spend more money for higher quality items that they'll expect to last for a long time.
Like spending $12K on a device, but knowing it will be awesome for the next 15~20 years.
The hopeful analogy presented by the submitter of the article "Exactly the way the collapse of the vacuum tube industry killed the economy, I hope." is misleading. Vacuum tubes were supplanted by the rise of a technology, solid state transistors, that performed the same function as vacuum tubes but was superior in every way and enabled significant advancements in electronics and computing.
Sixty years on, and there is no obvious successor to solid state transistors that can promise the same kind of exponential boost in capabilities that was seen when Shockley et al. invented the first BJT. A breakthrough may occur, but it is also possible that no such breakthrough will occur. Imagine that the transistor was never invented and that advances were still measured by incremental improvements in vacuum tube technology. The personal computer of today would be the size of a refrigerator and have the computing power of a graphing calculator. This is the premise to consider when invoking the possible demise of Moore's law. Imagine if your great grandchild's personal computers are only 20 times faster than today's computers rather than one million times faster.
Once CPU have reached their absolute limit, you'll just have to wait if you want a lot of computing done. Just like we used to when computers were first invented. The real limit to what can be done will most likely be communications issues bringing all the data together to be processed is likely to be a much slower process. And exactly what will your iPhone need to be doing with all that super computing power anyway?
Having survived the collapse of the vacuum tube and then the transistor/solid state and then the IC based economies. All I got to say about that is; what you got to do is pay attention. I gotta admit that it get's harder as you get older..
When you can't scale vertically any further, you scale horizontally. What we need is software that will take advantage of multiple cores and hyper-threading in those cores. There are some out there. (Video editing comes to mind.) If we get started now, we will be all set when the end-game becomes imperative. That way, we can see processing improvement in the same relative footprint.
My wife doesn't listen to me either...
And even if it at first looked like a planetary alignment, that could just be orbital wobble.
________
Entranced by anime since late summer 2001 and loving it ^_^
I read (& posted a comment to) the Salon.com article, and I'll rephrase it here.
At some point, there should be a point where Moore's Law collides with the "law" of diminishing returns.
As others have posted here before me, we have seen a flattening of CPU speeds for a few years, and when's the last time you heard someone say, "I want a new computer, but if I just wait 6 more months, it will be faster, better, cheaper"? Indeed, the same has applied to laptops for a fair while now.
Sure there's room for Moore's Law to rule the mobile market for a few years to come, but once you have IBM's Watson in your pocket, how much more computing power is needed?
And yes, I modded my AMD 64 so it only has 640Kb, because that's enough for anyone - harumph!
--
Salon Kill File: Better letter reading on Salon.com.
http://salon.maow.net/
https://salon.maow.net/
Does that mean that the next version of Windows won't ever be able to run on any machine?
Who would win this election: Andrew Weiner vs Andrew Weiner's weiner.
Every time someone has looked to the future daring to assert technology will be disruptive to labor has so far been proven wrong.
Yet I still foolishly choose to ignore history in thinking this can't last forever... All bets are off as machines get smarter. God help us when they become smarter than us.
On my Commodore 64 all I need to run are on hundreds of disks with thousands of bootleg applications and games plus 16 color porn!
That's all I'll need in 20 years right?
Oddly enough, this wittering about Moore's Law breaking down Real Soon Now is somehow not a repeat from, oh, I dunno, EVER.
Seriously, was there ever a year when someone wasn't ranting on a street corner that Technology Cannot Progress Further and The World Is Doomed?
Who cares about transistor density in this day and age? Mass-production of multi-core machines is fairly recent. We're still learning how to take advantage of this. What I'm basically saying is that algorithm development is far more important right now.
They can make the same processors for 20-30 years and still convince people it's something new
Yep rubbish. and here's why.
More's law is actually about the economics of computing - computing power per dollar.
When AI and Robotics Really hit the fan, the cost of labour becomes Zero.
Making chips becomes Free, so computing power per dollar becomes Infinite.
A few decades after that we will have finished converting as much as possible (allowing for continental drift constraints) of earth's crust into computronium (at zero $ cost).
even if that was only 10 times denser, and 10 times faster computationally than today, we would have frakking god like computational power on tap.
Although I like Kaku as a scientist in general, he's not exactly immune to mythbusters-style "foot in the mouth" science.
One major thing he overlooks, is the high likelihood of cloud computing eventually taking over the role of the "processor" in most PCs well before then. This isn't just a fad technology that'll go away in a few months, it's probably going to be next evolution in computing since the introduction of the world wide web. Not only will it take over processing tasks, it'll also change the very way software is distributed by letting companies post a "master" copy of a program onto a cloud server, and then rent out usage time for an instance of the software in a user's cloud space. This way, the developer doesn't have to waste months of development time trying to track down bugs specific to different system configurations. This would alllow the developer to focus solely on the software's performance within the cloud, only requiring updates to the "master" copy as they are needed. The user would never need to worry about all the downsides to installing software, such as invasive DRM, software incompatibilities or malware, as the software would never actually be running locally on their system.
Likewise, processing power would also be rented out. A larger portion of CPU time could be purchased for an extra fee, on a sliding scale. One cloud computing becomes as flexible as that, one only needs the right version of client software to access their cloud vm interface and you could theoretically access it from any machine with enough local horsepower to display a window of a stream viewport of the user's workspace. (Probably any system from 1999 to the present)
After that, processing power becomes largely irrelevant unless you are working on something seriously data intensive beyond anything we can probably comprehend now.
8==8 Bones 8==8
From Wikipedia:
The Great Moore's Law Compensator (TGMLC), generally referred to as bloat, and also known as Wirth's law, is the principle that successive generations of computer software acquire enough bloat to offset the performance gains predicted by Moore's Law. In a 2008 article in InfoWorld, Randall C. Kennedy,[34] formerly of Intel, introduces this term using successive versions of Microsoft Office between the year 2000 and 2007 as his premise. Despite the gains in computational performance during this time period according to Moore's law, Office 2007 performed the same task at half the speed on a prototypical year 2007 computer as compared to Office 2000 on a year 2000 computer.
- I'll be happy once we finally break into quantum computers and can't have smaller transistors, and can't develop smaller or faster things than electron states, hopefully we'll start to focus on improving code, removing bloat, and making better software (madly optimistic, I know).
Two things we'll have for sure in twenty years (because that's a long time .... cough)
Affordable high efficiency solar, and fusion power, which by my recollection have both been stuck in a shifting and endless "next 20 years" prediction cycle for about the last 50 years.
A couple things you won't see in the next twenty years: Mass produced affordable flying cars (which totally pisses me off) and bacon in a squeeze bottle. The promise of never ending technological advancement is merely a diversion to keep the tech crowd from becoming focused, despondent and suicidal over the lack of same.
Enjoy.
This ain't no upwardly mobile freeway This is the road to hell
This morning I was waiting for the ferry on a barge-dock. The barge was rocking a bit in the bay. Two ripples were playing across a puddle on that barge but at two different frequencies. They passed through each other with virtually no interaction. Speaking as a non practicing EE - I thought about how much more there is to do at the gate level with solid state physics.
3-d, efficiency (power density, not volume density), photonics and plasmonics for non-interfering signals, its all at its infancy. Plasmonics in particular has huge potential because it gets many of the benefits of photonics while staying within more traditional solid state physics.
When transistors hit their limit, there will still be huge amounts of transforming to do. Even within technology, there are things to do: there's a whole avenue of domain-specific chips to pursue. With the exception of GPUs (and possibly cryptography), there has been until now no point in making chips to do one specific thing; by the time you made it, Intel's CPUs would be more powerful at doing whatever it was you were going to do anyway. When we really hit the limit of silicon, that will become a rich avenue to explore. epayroll