Notes on Gordon Moore

Since I posted a piece about Gordon Moore's transistor density prediction a while ago I have been wondering if Moore had real insight or just came up with a lucky guess. I don’t think it's really possible to answer that question but I did think that I could dig into his background a bit.

Moore rates his own spot on Wikipedia: http://en.wikipedia.org/wiki/Gordon_Moore
His educational background is in chemistry and physics - a good combination for someone interested in semiconductors. It's interesting that his career turned more into management than engineering. Most engineers don’t make that leap successfully.

He made his prediction in April 1965 which is 11 years after he got his PhD from Caltech. He had spent the entire time in the semiconductor industry in various key roles that led to the creation of both Fairchild Semiconductor and Intel. From these facts alone we might conclude that Moore had some pretty good insight into what semiconductors would sell well. Moore wrote a short piece about his own experience in becoming a manager of a new business: http://nobelprize.org/nobel_prizes/physics/articles/moore/index.html which is well worth reading because Moore is quite frank about how some of the business developments were more accidents than planned events. His focus had been as part of a small lab (at Fairchild). It just ended up as a very big company (Intel) partly because they were on the very forefront of integrated circuit (IC) development and partly because there was a great demand for what ICs had to offer. Moore describes how they tried different technologies at Intel, hoping to exploit one or all of them. Eventually one technology of the three resulted in a real product - the Intel 4004 microprocessor. This was a major step forward in the industry. Intel sold directly. They didn't use distributors as they had at Fairchild. Moore and his colleagues had learned a good lesson there and this helped Intel to be more profitable. The rest of the microprocessor story is history.

I rate Moore highly on a lot of different fronts. He's obviously a bright guy and had a good education. He learned well from past mistakes and kept an open mind along the way. These are traits that a lot more managers could do with. He also tried to keep things simple, by limiting bureaucracy and giving his engineers simple ways to get things done. A lot of this is common sense from some perspective but not all that many people seem to be able to keep their common sense when they are running a very large organization.

So does any of this help us to know if Moore had some innate ability to predict the future of the semiconductor diffusion industry? I don’t think so. I think Moore was vectoring from the current state of affairs and riding the top of a big wave. He was looking optimistically at the future from a happy vantage point. In the end, I think he got lucky. There was so much drive to build denser semiconductors because of the emerging markets that the chances are we'd have made Moore's curve or better. It is still a bit surprising that the curve is still valid today. Few things grow with that much stability. As my friend Jason observed recently, people were busting their butts to make sure the industry followed Moore's Law.
Nigel