Magic + Loss

Share this post
Is Moore’s Law Really Dead?
virginiaheffernan.substack.com

Is Moore’s Law Really Dead?

Virginia Heffernan
Nov 22, 2022
8
2
Share this post
Is Moore’s Law Really Dead?
virginiaheffernan.substack.com
The words “time” “costs” and “production” being crushed in a press
Wired illustration/photo SHAWN MICHAEL JONES.

OHM’S LAW (V = IR) states that the voltage across a conductor is proportional to the current flowing through it. Hooke’s law (Fs = -kx) states that the force needed to extend or compress a spring by some distance is proportional to that distance. Moore’s law states—

Well, that one doesn’t state. It wagers. It hazards a guess. It contains no constants, no special functions, no variables, no equations at all. Movie directors will find in Moore’s law nothing like the kind of pretty runes that John Nash or Ben Affleck might scrawl on a window with a wax pencil. In fact, Moore’s law is less a law than a flier, a bit of Johnson-era bookmaking. Every year (or two) for a decade—or so the “law” goes—engineers would maybe, probably, double the number of transistors they could stuff onto a silicon chip.

Such was the bet of Gordon Moore, then the research director at the illustrious Fairchild Semiconductor, an outfit based in Sunnyvale, California, that in those days was known mostly as a division of the Fairchild Camera and Instrument Corporation. Electronics, a throwaway circular for the radio industry, picked up one of Moore’s reports for Fairchild and published it as “Cramming More Components Onto Integrated Circuits” on April 19, 1965.

It was a hit. Prescient is too weak a word: “Integrated circuits will lead to such wonders as home computers, or at least terminals connected to a central computer, automatic controls for automobiles, and personal portable communications equipment.” Such wonders. The essay dates to a time when a computer costs $18,000 (around $170,000 today). To make one affordable for a household, let alone an individual, will take a miracle. Moore, who later cofounded Intel, gives the miracle a shove with his resounding vote of confidence in tech workers for whom “You can do it” turns into “You must do it” turns into “You will do it” turns into “It’s natural law.”

“Cramming” in this way sweeps you up in a tide of inevitability. It’s only slightly too grand to say the essay recalls the Declaration of Independence. (“When in the course of human events” also makes revolution sound like natural law.) For a piece about integrated circuits, moreover, “Cramming” partakes of the marvelous-fantastic genre; nothing Moore predicts for the future—and that future is now—is anything less than riotously fun.

Five years after the essay’s publication, Carver Mead, a renowned scientist who just this year won a tidy 100 million yen as the Kyoto Prize laureate, playfully dubbed its projections a law. Thus the standing orders to America’s postwar engineers became clear: Miniaturize transistors and reduce costs. A world of workers have applied themselves to this task ever since. As much as Moore’s skills as a would-be legislator, his infectious enthusiasm for semiconductors has inspired these decades of engineering commitment. “Cramming” is still well worth reading, especially if all you know of Moore’s law is the doubling thing. Not only is it vatic, robust, and refined at once, it’s a piece of prose without precedent, a speech act that brought into being a whole global economic sector.

Keep reading here at Wired.

Magic + Loss is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

2
Share this post
Is Moore’s Law Really Dead?
virginiaheffernan.substack.com
2 Comments
Stefan Adler
Nov 27, 2022Liked by Virginia Heffernan

Re: "In science, laws obtain. In tech, it’s mostly faith."

10ish years ago I was at a social thing talking with a startup CEO with either a Masters or PhD in particle physics. He insisted that Moore's Law was an actual law. I compactly (a rarity for me) pointed out that Gordon Moore had adjusted the schedule from 18 months to 24 months, the improvements were driven by R&D spend, and likely the speed at which the market could productively integrate change, so no, it was not a law and would eventually reach some limit. So a guy (I know, it shocks) whose letters were dependent upon actual laws of the universe could not tell a rule of thumb from an actual physical law.

Expand full comment
ReplyCollapse
Nancy Heffernan
Nov 25, 2022Liked by Virginia Heffernan

Moore’s Law was news to me, thanks.

Expand full comment
ReplyCollapse
TopNewCommunity

No posts

Ready for more?

© 2023 Virginia Heffernan
Privacy ∙ Terms ∙ Collection notice
Start WritingGet the app
Substack is the home for great writing