MacGougan at Large

Notes on Laws That Aren't Really Laws

Moore’s Law - The Power of Multiplication

Gordon Moore was a co-founder of the chip-maker Intel. In 1975, he noted that the number of circuits that could be fit onto a computer chip was doubling every two years. He didn’t actually predict that this would continue to be true forever, but it’s continued to be true for the fifty years since he pointed it out, and this pattern of accelerating growth has long since been deemed Moore’s Law.

Logic dictates that Moore’s Law can’t continue to hold true indefinitely. Eventually, it runs up against the Wheat Grains on the Chessboard Problem. According to an ancient legend, a savvy commoner, asked by a ruler to name his price for some service he’d performed, suggested that he be paid according to a chessboard - with one grain of wheat for the first square, two grains for the second square, four grains for the third square, and so on through all 64 squares. Surely that couldn’t add up to more grain than the ruler could afford to pay!

In fact, it ends up being more wheat than the entire world produces in a year. Such is the power of repetitive multiplication. “The most powerful force in the universe is compound interest,” is a quote that’s widely, but inaccurately, attributed to Albert Einstein. Although it might actually be true, even if Einstein didn’t say it.

At some point, Moore’s Law would dictate that computer chips should contain more transistors than there are atoms in the universe, which probably isn’t going to happen. Experts suggest that transistor density on chips is going to start leveling off over the next decade, but that scientists will find other ways to continue expanding computing power.

Personally, I’m ready for some leveling off of technology. Compound growth rates are great for investment funds, but dangerous for everything else. In biology, continually accelerating growth is pretty much the definition of cancer.

PS - Here’s a serious note on AI inspired by those civic-minded Norwegians. In 1990, when the huge oil deposit under the North Sea was starting to be tapped, they decided that - while the oil companies needed to be profitable enough to justify the large investment needed to extract the oil - the oil companies didn’t deserve all of the profit from this public resource. They negotiated a healthy chunk of the profits to go into a Sovereign Wealth Fund, to be used for the benefit of the Norwegian people. That fund is currently worth about $2 Trillion with a T. (That’s a bit more than $350,000 per Norwegian.)

The AI engines being built today learn language from our communications and are trained to do tasks by our prompts and feedback. Yes, those who mine the data should earn a profit, but I don’t believe they deserve all the profit that will be extracted from a resource that ultimately belongs to all of us. Let’s negotiate for a share of AI profits to go into an American or maybe international Sovereign Wealth Fund.