“No law is more powerful or important in Silicon Valley than Moore’s Law — the simple idea that transistor density is continually increasing which means computing power goes up just as costs and energy consumption go down,” Bob Cringely writes for I, Cringely. “It’s a clever idea we rightly attribute to Gordon Moore.”
“The power lies in the Law’s predictability,” Cringely writes. “There’s no other trillion dollar business where you can look down the road and have a pretty clear idea what you’ll get. Moore’s Law lets us take chances on the future and generally get away with them.”
“But what happens when you break Moore’s Law? That’s what I have been thinking about lately,” Cringely writes. “That’s when destinies change.”
Read more in the full article here.
not less is more, Moore is more.
“That’s when destinies change”
In favour of the user experience innovators.
“Going Forward” of course.
Nobody can predict where Moore’s law is going. With space running out, chips are now going vertical. Chips have to be pictured as 3-dimensional devices. More like a city, instead of a prairie Plus with quantum computing, bio-computing, and other new technologies, who knows where we’ll be in 10 years? The journey is the reward.
exciting stuff.
“There’s no other trillion dollar business where you can look down the road and have a pretty clear idea what you’ll get.”
BS.
All maturing manufacturing industries have product maps and performance curves that asymtotically reach a predictable pace of change.
Moore’s theorem has done little to predict processor changes (Power PC architecture losing out to Intel architecture, Intel’s mobile architecture losing to ARM architecture, and so forth). Moreover, buyers of computers do not calculate transistor capability at all when they buy. Why would they? It’s a meaningless number. What matters most to the typical user is responsiveness — i.e., rapid operations. Hence total system and OS integration trumps the number of transistors any given machine might have. Most consumers can find the clock speed of their processor or RAM attached to the motherboard, but that’s about it. These are weak proxies for system performance, but are better indicators than Moore’s unsustainable theorem.
What do you expect? His nickname is Dr. Doom. Frankly, this reads like an unqualified endorsement considering the source.
Wrong article.
“Hey, Rocky, watch me pulla rabbit outta my hat!”
ROAR!!!
“oops, wrong hat.”
Moore’s “Law” was only a guess, based on trend information from 1965, that for transistor density, dD/dt = k.D with k = 1/18mo . It has no underpinnings in basic physics or chemistry. It’s not a law, just a self-reinforcing trend that could break down at any time due to actual laws of physics and chemistry.
Actually, the original quote stated 12 months. Moore revised it later to 24 months. 18 months was created by the media by averaging the two numbers.
The real “observation” (not law) is that the number of transistors will double every 24 months.
People know it isn’t a law, just a prediction that has had an eerie ability to be accurate long after the original author ever thought it would be. That is why, half jokingly, people keep calling it a law.
I thought Cringely had retired.
I’d rather someone coherent pondered the questions of the universe and transistor density.