The Chip Insider®

Finding meaning in Gordon Moore’s passing: By now, I’m sure you’ve read many obituaries about his life and accomplishments. The volume of articles that exploded over the weekend after the Friday he died on March 24 was unparalleled. Some of them were a full page. It arguably topped the coverage given to Steve Jobs’ death, even though Gordon was far less seen in the public eye.

If you are a true insider, his accomplishments meant far more to you and your career than any of those articles could come close to revealing. The breadth of this across the industry could be witnessed, as internet portals clogged with an unparalleled level of grief posts bemoaning Gordon’s death and what his life had meant to them.

The bigger question is not the meaning of Gordon’s life. It is the new meaning we can find to continue after his death.

While Sputnik was the earliest catalyst that attracted technologists to semiconductors, Gordon’s 1965 “Moore’s Law” paper inspired us to seek ever-greater levels of innovation.

We came from the farthest reaches of the globe. We – not he – transformed the ‘Valley of Hearts Delight’ into Silicon Valley. Inspired, we took them from several components-per-chip to billions, as Gordon inspired the people he met with the oft repeated, “Just remember, whatever has been done, can be outdone!”

This is Moore’s wisdom of Moore’s Law and life in general: whatever you have done, will be outdone. It is your decision: Will it be someone else or will it be you doing the outdoing? The paired WildPhoton below speaks to this wisdom: after each sunset, there will be a sunrise. It’s up to you to make the most of each day. It’s a Kona sunset. I took it near where Gordon lived his last days, at the time thinking it so emblematic of how he lived his life, relaxing after the last node, but looking forward to the next.

Shortly after he retired, I asked him what he wanted to do next. Classic Gordon, he said, “I just want to come back in a hundred years to see how it all turned out.” There lies one of the meanings for us: Outdo what we have done to make the world a better place. Just think of the millions of lives saved as chips with billions of transistors became the AI engines to bring the first COVID vaccines to the world in less than a year. But I don’t need to tell you more about the impact of semiconductors, because you are an insider.

There is another meaning we can extract from Gordon’s death. It has to do with how we view and use Moore’s Law. We can use it for the good of the industry or for personal gain. Our industry is divided about it across multiple dimensions. Some are 1965 paper literalists, limiting it to density doubling.

Others expand it to include his 1975 paper’s clock of every two years, but then ignore the fact that if it changed once, it might have changed multiple times.

Others – such as myself – include his point about areal manufacturing costs not rising. Hence, cost-per-transistor halves every node.

Others expand beyond what’s in the paper. Such as seeing it as an exponential growth in computing power. Or by limiting it to logic only, ignoring Gordon’s use of memory chips to make his point.

Then there are those who proclaim “Moore’s Law is Dead” for personal gain, using the attraction power of its shock value. Notably, there’s Jensen Huang, who arbitrarily limits the density doubling definition to CPUs, which peaked long ago, to promote the continued density doubling of GPUs. I cannot say if Jensen uses it to promote Nvidia’s amazing accomplishments or never took the time to critically read Gordon’s papers, because GPUs were never included – nor did they even exist when the papers were written. There is even the YouTuber @MooresLawIsDead, who has managed to lever words to 156K subscribers and 28M views.

Of course, none of this can go without mention of Gordon’s own beliefs that the mechanisms he defined as driving the process must eventually end. He brings this up as early as in his 1965 paper and in virtually every paper on the subject after that. As for personal gain, Gordon was always embarrassed by his so-called “law” and never took ownership of it because it’s not a law in a technical sense like Ampere’s Law. It’s not even a principle or theory by these standards. In fact, Gordon may have innocently done the most damage with his VMIC 1993 paper titled, “No Exponential is Forever.” In that paper, Gordon laid out real physical limits to gate oxide thicknesses that would stop the industry at a quarter micron unless they were overcome.

Insiders can debate Moore’s Law without causing damage to the industry as long as it stays in the industry. We can lay out “red brick walls” and tear them down as long as the outside world doesn’t hear about them. It’s like the saying, “What happens in Las Vegas, stays in Las Vegas.” Problems only arise when the story leaks out of Las Vegas. That’s what arguably happened after Gordon’s 1993 paper due to his stature.

The problem is that chip outsiders see “Moore’s Law” as synonymous with “Innovation in Semiconductors.” When they hear an insider proclaim, “Moore’s Law is Dead,” they mentally translate it to mean, “Innovation in Semiconductors is Dead!” In the second half of the nineties, the semiconductor industry began to have a hard time attracting talent. I remember reading an article in an airline magazine that “semiconductors were the new rust belt of industry.” With the arrival of the internet, the best and the brightest wanted to study computer science or finance. The best jobs in the future would be in software code, not silicon chips. This accelerated after that, has progressed across the world, and today has even infected Taiwan’s best and brightest as a recent Digitimes article complained of a shortage of fab-oriented graduates there.

Blaming it on Gordon’s 1993 VMIC paper may well be correlation without causation. I can’t help but be reminded that when imec came out with similar proclamations about the end of Moore’s Law starting at a SEMICON West in the 2010s. Soon after, people began to ask, “Why do we need a research organization if innovation is stopping.” The story quickly changed at imec and today they see Moore’s Law as continuing into the future with an amazing series of new transistor technologies to make it possible.

Ten years after his VMIC paper, Gordon came out again with a new title: “No Exponential is Forever – But Forever can be Delayed” at ISSCC. Gordon challenged technologists with the call to innovate, “No physical quantity can continue to change exponentially forever… Your job is delaying forever.”

Interestingly, while Moore’s Law is violently debated, few ever ask the man who coined the term: Carver Mead. In 2015 during the 50th anniversary of the paper, Rachel Courtland (then at spectrum.ieee.org) interviewed him. Here’s what he had to say about Moore’s Law:

“This is a law [of] the way that humans are. In order for anything to evolve like our semiconductor technology has evolved, it takes an enormous amount of creative effort by a large number of smart people. They have to believe that effort is going to result in a successful thing or they won’t put the effort in. That belief that it’s possible to do this thing is what causes the thing to happen.

The Moore’s Law thing is really about people’s belief in the future and their willingness to put energy into causing that thing to come about. It’s a marvelous statement about humanity.” – Carver Mead

Note that Carver did not say anything about density, scaling, cost, compute performance-per-watt, CPUs, or GPUs. Carver’s focus is on how innovation occurs across the industry: Semiconductor Innovation is fundamentally different because it involves hundreds of companies with thousands of people, solving millions of problems while still hitting each node together linked in time. Innovation in semiconductors is diametrically opposed to “great man theory” where innovation came from individuals like a Newton, Edison, or Zukerberg.

Carver Mead, who coined the term “Moore’s Law” and sold his ideas about how important Gordon’s 1965 paper was to the industry saw it as a way to spur innovation. While Carver never said this, technically “Moore’s Law” is a chaotic attractor that leads to Emergent Behavior creating a Virtuous Circle. Carver’s read on Moore’s Law is the same as industry outsiders: It is synonymous with “Innovation in Semiconductors.”

So the meaning industry insiders can glean from Gordon’s passing is …

• To learn to accept Moore’s Law for what it means to outsiders: innovation
  • That it’s the semiconductor industry’s unique brand for innovation.
• And to mirror insiders who say it’s dead by asking, “So you’re saying you’ve stopped innovating?”

For those grieving insiders who ask, “Why does it matter?” Answer that it’s about attracting the best and the brightest to come to our industry so they can find new ways to keep the innovation Gordon ignited going. He would want us to. We can’t continue to make the world a better place for the next 100 years unless we attract talent. Semiconductors are engaged on the front lines of two great battles: re-globalization and climate change. To borrow Abraham Lincoln’s Gettysburg Address, “It is for us the living, rather, to be dedicated here to the unfinished work … to be here dedicated to the great task remaining before us...”