"Moore's Law is not a Law. It is a trend. He was charting a trend in the stock market. This is not a law." --L505
"It wasn't something you expected to join the archives," Moore said in a recent gathering with reporters. "I didn't think it would be especially accurate." --Cnet
"Demand for greater power efficiency in computing is on the rise from desktop to laptop PCs. With an Intel Core Duo processor, you get a balance of great dual-core computing capabilities and power savings. Its enhanced voltage efficiency supports cooler and quieter system designs as compared to traditional desktop and laptop PCs." --Intel
Hmm.. above sounds like something I've been saying for the past 5 years. We've seen this in the automobile industry when the V8 engine and the heavy 5000 pound american crap cans phased out too. What else did they work on? Making cars quieter and increasing efficiency in exhaust, air conditioning, heating, insulation etc. They worked on making the engines less polluting too.
History repeats itself. It's funny because lately I've been wondering why the same 1.6Ghz laptops have been for sale for the past few years.
The answer is that either
- Intel finally got a clue and started enhancing the power savings and efficiency of computers rather than accepting the typical lies propagated by Java/.Net/Ruby folks such as 'gimme the 60Ghz computer for my virtual machine to run smoothly because computers are getting faster every minute and no need to worry about memory allocation or bloat any more, and a single program should use all the memory available and all the processing power available otherwise it is just wasted'.
(You should put your pedal to the floor when driving a V8 Mustang or Camaro and keep it down to the floor all the time even when going around tight turns and even when you don't have money for gas and even when a cop is on your tail and even when you have passengers in the car who are throwing up on your seats and getting bruises on their body because otherwise that gas and horsepower is just going to waste. If you can't use the gas and the power what is the point of having it, I mean come on let's go and hit some deer on the road for eff sake.)
- Intel didn't get a clue and is just pulling this efficiency garbage marketing hype because they are hiding the fact that Moore's law is bull shit and they can't keep up with it. By changing their marketing over to efficiency instead of performance they can keep it a secret that their processors have crappy performance and that no one needs to buy their latest and greatest since our current computers purchased 5 years ago are just as good but take up a tiny bit more electricty.
- a combination of both of the above
Moore's law bet that things (might, maybe, if, could) stay consistent. Since humans and nature are inconsistent, and since computing relies on humans and nature, we can immediately tell that his law is flawed. It's not even a law. People just call it one because they bend it around into a "law" just for the sake of making it sound "official", when in fact it is a complete IF, BUT, AND, COULD statement from Moore.
The bible is true in 20 years, if, but, and could be, maybe, if you read it this way and make exceptions and excuses here and there, and yes yes yes that's it. It's a law. THE BIBLE IS THE LAW. JESUS IS MY MOM.
Moore's law is like a Casino machine that happens to pump out money every 200 coins for a while, or a Stock market trend that is correct for a short period of time and then all of a sudden the bubble bursts. Moore's law is a small peak, a bubble which is increasing in size consistently. All bubbles burst or go through some drastic changes. It may even get better than Moore's law - it may get worse than his law. But nothing stays consistent like Moore's law predicts.
What Moore did do, is protect himself by stating "if, but, and kind of, not really, sort of, maybe, possibly". That is not a law.
Moore didn't have any law. His statement was just a "if but and, however only for ten years, but not sure about it, and if, but, and, then it doesn't apply".
This is not a law. This is an iffy butty andy statement from Moore.
The complexity for minimum component costs has increased at a rate of roughly a factor of two per year (see graph on next page). Certainly over the short term this rate can be expected to continue, if not to increase. Over the longer term, the rate of increase is a bit more uncertain, although there is no reason to believe it will not remain nearly constant for at least 10 years. That means by 1975, the number of components per integrated circuit for minimum cost will be 65,000. --Moore
Did Moore forget about History
Moore's law is funny because it doesn't take into consideration the years that we did not have computers. We went from zero computers to lots of computers.
In 1920 or 1820 no one had a desktop computer. There was a big flat point.. in year 1500 we didn't have computers. In year 1600 we didn't have computers. So Moore's law is already incorrect if we look at history. All of a sudden computers were invented and then we have a big peak in 1970-2007. Obviously when something is first invented you have a peak - and if 500 people make a guess at what this peak might be, one of the guys is going to be close to right. If 500 people make a guess about how many goals an athlete is going to score in 5 games then one of the guys will guess close to spot on. Does this mean we can take this one guess seriously, even if this one guess was based on some historical charting.
This reminds me of stock market experts who make one correct guess about a stock for 5 years but then all of a sudden the stock market starts playing games - such as maybe we go into a war time or great depression - and then all of the Moore believers start making excuses such as "yeah but moore was right except for this period here, and I read bible messages backwards and you can only focus on this message since this message is right and it happened". Something will totally screw up our computer pace trend such as:
- everyone being in the army and at war
- some other technology will be more important than transistors such as insect sized airplanes that can deliver mail physically rather than digitally
- computers will suck up so much electricity when they are 500Ghz that we will have to invent something more efficient than the current transistor anyway, just like we got rid of vacuum tubes since they were so big.
Oil isn't cheap, that is why cars improve on mileage/pollution control rather than HORSE POWER these days. Going Fast isn't the only thing. Moore's law encourages bloat ware and it encourages people to think that no longer do programming laws apply, such as wiping your bum after allocating 500GB of memory.
Moore's Law is the same attitude back when Gas was cheap and people drove their V8 gas guzzling huge automobiles around town:
- Gas is cheap, No Worries!
- HorsePower will double each month!
- The Cost of Gas will decrease 90 percent each year!
- Humans will be 24 feet tall in 200 years!
- In year 2000, the world will end!
It is stupid, like this car below:
Forget about the cost of Gas! Cars horsepower will double each year and gas will get cheaper!
Forget about memory and CPU usage, electricity is cheap and transistor etching is allowing us to double our computer speed and memory each day! Now we can run 50 JVM's (java virtual machines) and .NET runtimes without any worry! It is the exact same attitude the big V8 Gas Guzzling American Car people had.
Do you see a problem above? A big car is purchased (a boatmobile) and it is so big and so inneffecient with its big V8 engine in it.. and no one seems to care. Same thing is happening with computers. Bloatware is downloaded each day, people buy bigger hard drives, and everyone thinks we will be using 500GHz computer processors in the next few years.
What happened with cars? Now people drive around cars that get over twice the mileage of the big V8 boats that people purchased when gas was cheap. Nowadays you don't hear people saying stupid things like how Gas is cheap and how we don't need to worry about the mileage of our american boatmobile. You don't hear people saying that horsepower is doubling each day or year either. What do you need all this horsepower for? Why do people buy these cars:
What happens with computers? I don't think my Electric 1500 watt heater should suck up as much power as my 500Ghz computer.. so obviously something is going to change with computers. Laptops aren't going to have huge batteries in order to power their 500Ghz processors. We aren't going to be carrying around MainFrame computers in our brief cases.. so something is going to change or stay the same. This change, or this stay the same era will blow moores law out.
Where are the 80GHz laptops?
Right now (year 2007), laptops are having trouble keeping up with pace. For about 5 years now I've seen the same 1.3 or 1.8Ghz laptops for sale! Either we are going to settle down and continue to use computers in the 1-5Ghz range for YEARS, or we are going to invent some other technology not as related to current transistors (such as water based nano robot based computers instead of wire based).
It's not Ghz! It's Transistors!
Yes, GHZ does not actually directly relate to transistor count.. what Moore was talking about. However laptops today, whether or not we use GHZ as our reference point.. are not getting faster like they were. Transistor count, most people could care less about.. if the softwre doesn't run any faster.
Look at the typical idiot of year 2008 who goes out and buys a 64 bit parallel PC and realizes he just wasted his money on something that would be better in a server environment and has no impact on his general deskop work he does.
What is more than likely, is that we've hit a peak - and we probably won't have water based computers or anything really splendidly new for a long long time. What is clear, is that Moore's "trend" applied for a while and was just a trend, not a law.
Like I've said, I have seen the same 1-3Ghz laptops for sale for years now. Sure, transistor amounts on chips (what Moore's law is about) does not equal Ghz directly. But Ghz is an indication, a small indication of what is happening. We can look at other factors too, such as studying the transistor count... and we can claim that parallel processors will keep Moore's law going...
But parallel processors are useless for the desktop computer user today and if people actually buy them then they are fraudulently following Moore's law just for the sake of following his law.. even though there is no such thing as a Moore's law in the first place, since his statements were completely IF, BUT, AND, MAYBE, POSSIBLY.
Either the size of battery to power the hefty ghz and hefty transistor count is restricting companies like Intel from moving further, or the etching of transistors can no longer be improved with current technology, or a combination of both or.. etc. etc. etc.
Computers will no longer double in transistors like they used to - sorry. When you make a transistor the size of an Atom, do you split the atom and create an atomic bomb computer?
Then people start throwing out the excuses like "yes but Moore was right for that amount of time.. those 10 years that if we draw the charts a certain way just look right. And you can't count those years where the graph was wrong, you have to average it out in a certain way to make Moore look right, and ignore every year where he was wrong and after thirty years you can't pay attention to his law any more because laws don't apply then. You can only focus on this one little section where Moore was close to being right, even though he said "IF BUT AND MAYBE POSSIBLY KIND OF SORT OF" .
Just like I can only focus on "certain" messages that are found when the bible is read backwards, but not other messages which completely disqualify the bible backwards reading.
Consistent Computing Laws are Flaws
Since humans are inconsistent, and humans control computer manufacturing.. obviously Moore's law will not hold up. We will start focusing on reducing the power of computing significantly, or we may even go to a war causing us to focus less on computers for a certain time, or we will be hit by an asteroid, or we will start focusing on nanotechnology (or both) and/or transistors, or will be caught up in other technologies such as water based electricity rather than transistors. Humans are inconsistent and we cannot place our bets into consistency. Moore's law is a bet that things will remain consistant, which is total bull shit. If anything history has shown us that trends are extremely inconsistent, even if consistent for a period of time. Even if computers improve transistor count, we will have to ditch transistors at some point and come up with something like transistors but not actually transistors. Of course all the Moore people will be saying that "yes but those things are like transistors and they are now improving at 4 times every 1 year which is just like the transistors since these are double the power of transistors and just transistors in disguise".
Cars may have doubled in speed each year for a while or maybe they doubled in horsepower or mileage every 3 years or 6 years or 6 months..and someone may have just happened to guess that cars will go twice as fast every 10 years.. but we all know that this isn't really helpful since 500 people could have made guesses and only one guy got recorded as guessing. Eventually we will no longer be worried about horsepower or mileage since there are so many other things such as cargo space and fuel type and pollution amounts. Focusing on whether or not horsepower has increased twice each two years is focusing on the wrong thing. We are so focused on the transistor and not focused on other more important things. Steam engines phased out, dinosaurs phased out, transistors will phase out.
Transistors will no longer be important once they can't put transistors on a small area any more - they will have to invent other technologies. Moores law is like a law that says I can win X amount of money in the stock market based on stock market history. It may be true for a while and it may even be that some stock market experts can predict a stock will go up for 2 years. But bright intelligent stock market folks will NEVER tell you that the stock market will be consistent for XYZ amounts of years. That's why Moore used all the IF, BUT, AND, MAYBE, COULD BE's in his statement that people nonsensically call "a law".
Trends VS Laws
Never in history has a trend been consistent. Trends are not LAWS. Hurricanes are random, earthquakes are random, weather is random, humans may die just like the dinosaurs died.. so what makes "Moore's Law" something that we should take seriously if trends are never trends? It is not a Law. It is a trend. He was charting a trend in the stock market. This is not a law.
Something useful would be a prediction that told what we should replace transistors with, such as water based electricity transportation instead of wire based electricity transportation.. as all living things generally have a lot of water in them and are very efficient at transporting electricity. Focusing on how good transistors will be in X years is a waste of time if transistors aren't even that useful compared to say water based electricity.
Technology, not laws
If they can create a pinball machine small enough to see in a microscope, then they can invent simple bacteria like nanobots that move around and try to eat other stuff and get bigger. Training these nanobots or programming them to build efficient computers for us out of sugar and water will be the next step. Water is very important since it can transfer electricity much more efficiently than wires and circuit boards. Since we will be entering a water technology age, this will blow moores law out since we will have to learn about water, and micro machines, rather than transistors. Someone will create another law about nanotechnology improving at XX amount and he will guess it right or wrong just like folks predict the stock market based on previous evidence and some get it right. At some point, nanotechnology might just happend to improve at the same speed as Moore's law, and someone will find backwards messages in the bible and declare that the bible has information about nanotechnology in it. Those working on more important things like the micro pinball machines won't be noticed, but stupid folks that make ridiculous laws which don't actually matter will continually be noticed (statistics whores rather than inventors).
We can now magically call stock market analysers "lawmakers" too. Yes, all folks who have guessed the stock market right once in their life can now make "laws". The fact that computers are improving their transistor space matters as much as how much horsepower people have got out of V8 350 engines in 1975, folks.
More important things to worry about are new technologies that don't require V8 350 engines, and computers that don't require wires but use water to transport electricity instead.
Links to nano scale pinball machine:
Even if pinball is a fairly useless nano sized machine.. a small little car that eats sugar isn't as useless.. and if we can emulate bacteria we can then improve our nano machines more and more until computers with circuit boards are no longer interesting, and Moores law won't be measuring transistors any more - we will be inventing other laws such as the speed we can improve nano machines and the time humans have left on earth since some guy has created a nano machine that will come and kill us all in our sleep.
So instead of predicting how many transistors we can pack into a small area, let's start thinking about other technologies that may not even require transistors. Let us also start thinking about the future of humanity.
What I am saying here is that people bring up Moore's law as if it is something we should take seriously and actually spend time worrying about. It is not a law, it is a fad and a trend. Someone will just claim that his "law" is still correct if you take into consideration the fact that he iffed, butted, and anded the entire "law". Laws cannot be "if, but, anded". That is a trend. It's a fad. Moore's "Law" was really just a charting of the stock market.