Is Moore’s Law Going out of Style?

By now you are probably familiar with Moore’s Law, which is named for Intel co-founder Gordon Moore and which holds that computer chip power doubles every 18 months or so. What’s remarkable about this observation is that Moore made it in 1965, and it’s held up pretty well.

Yet Moore’s insight, however powerful, did not become widely known outside of technology circles until comparatively recently. According to a recent search, Moore’s Law was mentioned 66 times by general news sources in the Nexis database during the first six months of 1995. Looking at the number of citations of Moore’s Law in a series of six-month periods since then, it turns out that they doubled roughly every 18 months through much of the 1990s, spiking to 591 mentions in the first half of last year. By then, its incantation had become part of the soundtrack for the era of New Economy magical optimism that we’ve recently lived through.

In a column the other day, I looked at some of the manifestations of the New Economy that seem due for the full backlash treatment, and while I don’t think Moore’s Law itself is going to come under attack, I do think that general tolerance for visionary extrapolation has suffered a great deal. I also think that’s a good thing.

The obvious appeal of an insight like Moore’s is that it was bold and iconoclastic, and it turned out to be right. As you may have noticed in the past few years, a critical part of the New Economy pose was to make predictions that seemed bold and iconoclastic and to chastise all doubters as short-sighted fools who just didn’t get it. Sometimes these predictions were accompanied by a bar graph, showing two years of data and extrapolating another five or 10 years of fantastic growth. (Such a prop is what allows business-world pundits to refer to their predictions as “projections.”) Sometimes these predictions were accompanied by a knowing reference to Moore’s Law. Eventually, when enough self-styled iconoclasts were all saying variations on the same thing, what you were hearing wasn’t visionary; it was a blend of conventional wisdom and wishful thinking.

The big myth about ga-ga projections, a myth mostly promulgated by people who make them, is that they are hard to make. Not so. It’s actually very easy to look at a very small amount of data and project it endlessly and seamlessly into the future. In fact it’s so easy that it’s often meaningless. And now that the future has shown up, it turns out that while Internet usage grew quickly, that does not mean that one could extrapolate from there that, say, couch buying over the Internet would have caught on in quite the same way. What a surprise.

As for Moore’s Law, while I haven’t seen any evidence that the trend it describes is about to come to an end, what does seem to have lost momentum is the trendiness of dropping references to the law itself. In fact, if you’d projected that Moore’s Law citations were going to keep doubling every 18 months indefinitely, you would have been wrong. The peak seems to have coincided with the peak of New Economy optimism generally, and mentions actually declined mildly in the second half of last year to 572. Which proves what? Well, nothing much really, except that some projections are worth more than others. And that, actually, is not such a bad point at all.