Print
In a 1965 paper, Intel co-founder Gordon E. Moore predicted that the number of transistors on an integrated circuit would double approximately every two years. This prediction has proven to be uncannily accurate over the years and has come to be known as Moore’s Law. But it’s not going to hold true forever, is it? Well, it’s believed that like all things good, Moore’s Law too will come to an end one day. The question that remains, though, is when. Noted theoretical (and often theatrical) physicist Michio Kaku feels he has the answer.
Kaku believes the days of Moore’s Law are numbered, and foresees its demise in around 10 years or so. Like many others, he feels it’s becoming increasingly difficult for us to sustain this exponential growth using just traditional silicon-based computing. According to the theoretical physicist, who is a co-founder of the string field theory, even three-dimensional chips like the ones recently introduced by Intel aren’t going to be of much help due to problems like leakage and heat.
“So, what’s beyond silicon?” Kaku asks in a recent BigThink.com video. “There have been a number of proposals: protein computers, DNA computers, optical computers, quantum computers, molecular computers.”
“If I were to put money on the table, I would say that in the next ten years we’ll simply tweak Moore’s Law a bit with chip-like computers in three dimensions, but beyond that we may have to go to molecular computers and perhaps late in the 21st century quantum computers.”
Image Credit: Murketing
Comments are closed on this article
aarcane
May 02, 2012 at 7:21pm
While Moore's law as stated only applies to transister count, it's more generally applicable to overall performance. Anyone with the brains to figure out that a literal interpretation isn't always necessary or proper can see that Moore's law will continue to apply to new technologies both in transistors and in future transistor-less technology. Furthermore, while I do predict a few MAJOR leaps in defiance of moore's law, such as to Quantum Computing, I believe that Moore's law will still be applicable to newer technology.
It's already aparent that Moore's law applies to pretty much any unit that can be said to be atomic. It's recently begun to apply to processor cores and cache sizes instead of simply to bare transister count as well. Moore expressed his law in terms that applied at the time, that is all.
timmyw
May 02, 2012 at 11:57am
Clarke's First Law: "When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong."
whathuhitwasntme
May 01, 2012 at 5:53pm
you know this seems like the same old story over an over every couple of years.
every year they improve the process or come up with a new one
john867
May 01, 2012 at 10:31am
Oh really?
Ten years for Moore's Law eh?
Surely this guy is smart enough to know they've been saying exactly this continuously for 20 years. The end of Moore's Law has ALWAYS been 10 years away. Yet every year, the end keeps getting pushed back another year.
Performance is another matter. It seems we have reached a point of diminishing returns...more transistors is giving us ever smaller increases in perceived computing power.
digitaldoc77
May 01, 2012 at 6:45am
It does seem like Moore's Law is stuck lately. After all, neither Bulldozer, nor Ivy Bridge have offered any real big performance improvements forward (although Ivy Bridge is hopefully laying the groundwork for this, and an outside chance that Bulldozer did as well). Still, we all got used to linear progress over the last two decades, and as the limits if lithography get hit, and max out the number of cores, there will definitely need to be a shift from silicon to new substances. Not sure if proteins will ever be workable as a computer though.
LatiosXT
May 01, 2012 at 8:02am
Moore's law concerns the number of transistors we can fit in a given area, not performance.
JohnP
April 30, 2012 at 8:42pm
Geez, and not a word about graphene. It can be used as a semiconductor, it is one the best conductors or electricity AND it is one of the best conductors of heat. It has been made literally in sheets one atom thick. Its better than silicon is so many ways.
People trying to prophesy the future always fail as it is impossible to predict what they cannot possibly know in advance of new ideas. All they are really doing is taking what they DO know and extending it into the future. Take graphene, it was only a theoretical idea and no one believed it could be a real thing. Two scientists in 2004 came up with the idea of using scotch tape and a block of graphite (pencil lead is graphite) to see if they could tease out something. They won the 2010 Nobel prize for it. Scotch tape, pencil lead. Nobel prize.Now ask Kaku if he predicted graphene or what the NEXT break-thru will be.
compro01
May 01, 2012 at 7:17am
What about graphene? Unless individual graphene molecules are transistors, it's not going to help Moore's law keep going.
Moore's law is about transistor density. Not processing speed.
JohnP
April 30, 2012 at 8:43pm
And here IS the next break-thru. I dod not see this coming, how about you Kaku? I rest my case.
http://www.extremetech.com/computing/127887-silicene-discovered-single-layer-silicon-that-could-beat-graphene-to-market
theplustwo
May 01, 2012 at 7:39am
Well, I think his point still stands that the manufacturing process is eventually going to bottom out... Intel is predicting having a 4nm process by 2022 ( http://en.wikipedia.org/wiki/4_nanometer ). Exactly 10 years from now, as Kaku says in the video.
There's only so far you can go getting thinner and thinner until eventually you're working with single atoms, after which point quantum mechanics come into play as he says in the video.
compro01
May 02, 2012 at 3:56pm
Quantum mechanics are already in play. Quantum tunneling is already resulting in current leakage issues and it's only going to get worse the smaller you go.
Biceps
April 30, 2012 at 10:43pm
Yep, never say never! I was thinking the same thing about graphene - especially since it's been in the spotlight so much recently as a possible replacement for silicon.
People were predicting Moore's law would end a lot sooner than 10 years from now, and other advances have proved them wrong so far.
vrmlbasic
April 30, 2012 at 10:51pm
Yep, this physicist seems like someone who just wanted his 15 minutes of fame and was willing to use/risk his professional image to get it.
Even if he is right, how does he know that the law won't apply to whatever computational method we "evolve" to? If we jump into molecular, quantum, or DNA (reminds me of Jonny Quest) computers, couldn't Moore's law hold true with just a little tweaking of the language? Replace "transistor" with "cell" and "chip" with "petri dish", for instance? ;)
Penterax
April 30, 2012 at 8:11pm
I think he's right on the end if you are talking about current silicon-based technology, but I wonder how much the multiple core technology, which is really just starting out, can keep silicon going.
Kaku of course isn't the only genius around (he didn't imply he was, far from it I think), and there are some really smart teams working on silicon.
What I think is that there may be kind of a lag in the Moore's Law-type effect, and then we'll see a jump forward - one that may shatter the Moore's Law rate of progress. If we can actually develop quantum computing to the point where it is workable, it will make silicon look like chiseling on stone.
;)
