The news emerged yesterday that Gordon Moore, semiconductor pioneer, one of the founders of both Fairchild Semiconductor and Intel, and the originator of the famous Moore’s Law, has died. His continuing influence over all aspects of the technology which makes our hardware world cannot be overstated, and his legacy will remain with us for many decades to come.
A member of the so-called “Traitorous Eight” who left Shockley Semiconductor in 1957 to form Fairchild Semiconductor, he and his cohort laid the seeds for what became Silicon Valley and the numerous companies, technologies, and products which have flowed from that. His name is probably most familiar to us through “Moore’s Law,” the rate of semiconductor development he first postulated in 1965 and revisited a decade later, that establishes a doubling of integrated circuit component density every two years. It’s a law that has seemed near its end multiple times over the decades since, but successive advancements in semiconductor fabrication technology have arrived in time to maintain it. Whether it will continue to hold from the early 2020s onwards remains a hotly contested topic, but we’re guessing its days aren’t quite over yet.
Perhaps Silicon Valley doesn’t hold the place in might once have in the world of semiconductors, as Uber-for-cats app startups vie for attention and other semiconductor design hubs worldwide steal its thunder. But it’s difficult to find a piece of electronic technology, whether it was designed in Mountain View, Cambridge, Shenzhen, or wherever, that doesn’t have Gordon Moore and the rest of those Fairchild founders in its DNA somewhere. Our world is richer for their work, and that’s what we’ll remember Gordon Moore for.
You can read our thoughts on Moore’s famous law. If you ever wondered how Silicon Valley became the place for electronics, the story is probably much older than you think.
16 thoughts on “Gordon Moore, 1929 — 2023”
“Perhaps Silicon Valley doesn’t hold the place in might once have in the world of semiconductors, as Uber-for-cats app startups vie for attention and other semiconductor design hubs worldwide steal its thunder.”
As noted the US is primarily where the design aspect happens.
A long and brilliant life. RIP.
Moore’s law is in fact a very clever example of prefiguring reality by coining a term. As you said, many times it looked like the law was going to fail, but people innovated to keep up with it. Nothing really enforces Moore’s law except the fact that Moore said so, and we’ve benefited greatly from it.
It can also work the opposite way in favor of businesses revenues. If you follow Moore’s law, you can leverage your gains by artificially slowing the progress and selling new MCUs 2x faster every two years for 50 years rather then releasing 100x faster MCUs right away.
Moore’s law was just an observation. It was used by companies to target product launch dates so that they wouldn’t be behind the competition when the product hit the shelves.
Much like Ohm’s Law, after observing current and voltage I suppose…
Not really. Moore’s Law can be derailed in a post-apocalyptic situation, for example. Ohm’s Law will be here forever.
Moore’s law can be artificially kept, because there’s nothing stopping you from making a larger chip with more transistors, as long as you’re following the version that counts transistors per chip instead of per area. It’s also possible if you forget the price point of Moore’s original prediction and use more expensive processes to increase the transistor density.
You see, there is not one “Moore’s law”, everyone made up their own.
Moore wasn’t the only one to make the observation. My father, working with Grace Hopper, made a very similar observation 10 years earlier in regards to the processing power of computers, rather than the number of transistors on a chip. It helps to be head honcho at Intel when you make the observation rather than just some guy working at ONR.
I would think money has a part to play in the advancement of technology along with competition.
Moore’s law “held” because people kept re-defining the law, including Moore himself.
The original version was about transistors per area per dollar, at the optimum price point. Then it became transistor per area, transistor per chip, and Intel fudged it up with some sort of performance metrics at some point.
When silicon valley was actually about silicon…
Wish people would stop calling it a Law.
It is a cool marketing term, and makes for nice copy. It is an observation only, and there is nothing to say it will be followed.
We are now seeing effects of it in terms of “7nm nodes” which are really not 7 nm sized features but “7 NM, TM”, with the marketing hype making it all rather meaningless.
And wasn’t the original observation based on 18 months, and then got stretched out to 2 years?
It’s never been considered a law in the sense of the law of gravity; it’s just a law in the same way that Murphy’s or Godwin’s are laws: a pithy observation that is reinforced by the universe despite the absence of any particular reason for the universe to do so.
Well, it isn’t even reinforced by the universe because people kept changing the criteria of what exactly is doubling, and how long it takes.
That’s called the Texas Sharpshooter fallacy: you take six random shots at a wall, and then paint the target around the tightest cluster.
A refreshingly to-the-point Obit.
Another hero was lost to us…
Please be kind and respectful to help make the comments section excellent. (Comment Policy)