“A well-known theoretical physicist has taken direct aim at a key theory in the computer industry, saying Moore’s Law is collapsing,” Sharon Gaudin reports for Computerworld.
“Physicist Michio Kaku, a professor of theoretical physics at City University of New York, said in a videotaped interview on BigThink.com that time is running out on the 47-year-old law,” Gaudin reports. “‘In about 10 years or so, we will see the collapse of Moore’s Law,’ Kaku said. ‘In fact, we already see a slowing down of Moore’s Law. Computing power simply cannot maintain its rapid exponential rise using standard silicon technology.'”
Gaudin reports, “The prediction was made by Intel co-founder Gordon Moore in 1965. It holds that the number of transistors on a chip doubles about every two years and can be done inexpensively. Kaku, like so many scientists before him, said recently the two main problems that will derail Moore’s Law are heat and leakage. ‘That’s the reason why the age of silicon will eventually come to a close,’ he said. This is far from the first prediction that Moore’s Law is failing… Kaku said molecular computers hold promise but he sees ‘enormous problems’ with quantum computing and doesn’t expect it to really mature until the late 21st century.”
Read more in the full article here.
[Thanks to MacDailyNews Reader “Judge Bork” for the heads up.]
Why should Moore’s Law be restricted to the use of silicon technology?
It’ the law of large numbers. You know the one that says a company worth $500 B market cap can’t double earnings.
Not sure I need any more computing power. Guess I could have Google goggles with a wrist watch sized iPad controlled by Siri.
I dunno. I would like to see Handbrake encode a video in 10 seconds instead of 30-45 minutes with a standard iMac.
Here are the silicon independent limitations:
A) There is only so fast you can move an electron through material.
The solution to A is to make Chips and boards smaller so the electrons do not have to travel as far. This is fine and good until you hit…
B) An atoms radius is typical measured in angstroms (1 angstroms = 10 nanometer..yes that is the nano in nanotechnology). So how close and how small can you make Chips and boards. We are getting close to features only an atom or two in size.
just my $0.02
Umh, an Angström is 1 x EE-10 metres, i.e. 0.1 nanometres.
What an idiot. Does he not believe in innovation? Who said we will be still using silicon in twenty years? As industries grow, so does technology. I think he is stuck I his own narrow mindedness.
Actually, he is one of the most open-minded theoretical physicists out there.
I think the point is that the past and current progress in computing is based on continuing predictable advancement using the same type of technology. In this case, silicon-based technology.
Perhaps in twenty years (or 100 years), we will be using a significantly different technology for computers… But in that case, Moore’s Law probably does break down during the transition period between silicon-based computing and whatever follows due to “innovation.” It will be like going from vacuum tubes to transistors.
You are the narrow-minded idiot. Tell me O’ oracle of innovation, what science your assertion is based upon?
Kaku graduated summa cum laude from Harvard University in 1968 and was first in his physics class.
@ Gainstr: Considering Michio Kaku toenail clippings have at least double your IQ, I don’t think it’s wise to call him an “idiot”.
It’s a good thing that IQ does not necessarily have anything to do with success or your comments just might hurt my feelings then.
“Scientists study the world as it is; engineers create the world that has never been.” – Theodore von Karman
Which is to say, scientists are very good at predicting the past.
This is ridiculous; some scientists started predicting Moore’s Law collapse like thirty years ago.
I am sure that years from now in the future we will still be reading those scientific revelations about how Moore’s Law is expiring over and over again.
What it’ll do is stop the mind-numbing cycle of: making faster hardware and then immediately bloatwaring it so it doesn’t run any faster than the last gen hardware ran doing similar tasks. Hopefully we’ll go through an age where the key to speedy hardware will be extremely efficient software.
Yes, this. Exactly.
This is the single greatest comment I have ever read on this site.
It doesn’t matter how much faster devices get, somebody will bog you down with wasted processor cycles.
AMEN BROTHER! When I was still using Windows, THAT is the reason I stuck with Windows 3.1 for years after it was obsoleted. I could do an OS re-install in 3 minutes.
… prediction to silicone for a couple of reasons. Most importantly, that is where the “law” was expected to apply and where it HAS applied to date. Almost as important is the likelihood that some significant change/update to the “law” will allow the computer ‘verse to continue to expand, just with different measures.
He isn’t an idiot. He does believe in innovation. He is well aware that even “laws” must evolve or die. Or are you an ID Believer?
The profit motive will eventually make optical computers a reality.
And then…pundits in the news media will complain about the silicon fab plants of the world who were put out of business by the new upstarts.
Pathetic reporting by my view.
At this point there is no viable alternative to silicon technology. Granted the “Core” race is just beginning. When in doubt add more cores to the chip. More cores on a chip = more transistors on a chip.
There is rather simplistic answer :
Gallium arsenide. It’s vastly superior to silicon and manages frequencies up to 250 GHz, is more heat and noise resistant and performance is extreme. A bit more expensive and with a higher power consumption, though.
This guy is an expert in his field; how many here calling him an idiot are experts in the same area?
Enquiring minds want to know.
Time to start designing the Isolinear chips. 😉
Interesting. I was thinking of orthoparabolic chips providing 8 gigaquads of power, as the geeks of Start Trek would say.
In other news, Domino’s Pizza is offering something called “artisan” pizza – a mass-produced pizza that substitutes feta for romano. The definition of “artisan” is something hand-crafted and made in limited quantities. These days everyone’s jumping on the “artisan/artisanal” marketing bandwagon by sprinkling some feta or basil or some indeterminate crunchy grains on food and calling it “artisan.”
/rant
Look… any particular physicist can do dozens and dozens of pop science TV shows and personal appearances, but that doesn’t make them a media whore.
Especially when they say things calculated to get press.
Press attention. On themselves.
When they can no longer double the speed of the chips, then the price will have to be cut in half every two years.
We can call that . . . “Steve’s Law”
MDN has a high percentage of science-denying right wing commenters, what else would you expect?
Kaku graduated summa cum laude from Harvard University in 1968 and was first in his physics class. The guy built an Atom smasher for his High School Science project. He is infinitely smarter than the denier rubes represented here.
It would be laughable if it wasn’t so pathetic.
Two things not covered.
• Not every device will need to be doubled in speed. Some are fast enough already.
• AI … Artificial Intelligence is the next game changer. Sometimes it isn’t how fast you go it is intuitively knowing what to do along the way. Consider intelligence working even 100 or 1,000 times faster than a human mind assisting the person developing and innovating on something. Doesn’t need a break working 24/7 searching everywhere and running simulations showing possibilities and flagging the dead ends.
Storage will be the next wall after that if not before that.
Michio Kaku is not an idiot. He’s the opposite.
Anti-idiot?
Open minds are always ready for change and growth. In 1908, the Director of the U.S. Patent Office recommended to the President that the office be closed down because “everything that could be invented already has been.”
Haha,
These same people once said the world was flat and the sun rotated around the earth.
Sad, just sad.
That’s deep…
Moore’s Law is not a law of course. It is more of an amazing remark about the pasce of innovation and miniturizatrion in silicone chip design and manufacture. It says nothing about the possibilities it innovate in other ways to continue the growth in computing power. Multiple cores with programs optomized for them and opperating systems intelligent in dividing the work is what is happening now. Who know what will be next?
Exactly. Remember the kefuffle in the late 90s/early 2000s about the impending disaster for hard drive storage sizes, wherein data density could not get any higher? Then Boom! New alloys and the “crisis” evaporated.
Precisely. People are acting like Moore’s “Law” is akin to the First Law of Thermodynamics. Instead, it was simply an insightful observation of the trends of empirical data associated with processor development. The amazing thing is how well CPU progress actually tracked that model for several decades.
The media will attempt to sensationalize just about anything as a potential catastrophe.
No duh! This article is just plagiarized old news. Try graphene.
The answer is simple: Stop calling theories “law”.
Biztheory?
Sorry… Just had to.
And there in lies the problem with referring to anything that isn’t a physical constant or a legal document as a “law”.
It’s just as ridiculous to imbue Moore’s “Law” with any significance outside of it’s usefulness as a guideline as it is to talk about Apple’s growth as constrained by the imaginary “law of large numbers”.
Oh, I’m sure you’ve had experience with both Murphy’s Law and O’Toole’s Corollary (Murphy was an optimist).
Indeed, “Moore’s Law” might as be called “Geek’s Guess” – an extrapolation of an observation that has no theoretical underpinnings at all apart from the reasonableness of the growth following compound (41%/yr) rather than simple interest.