The basic premise of Moore’s Law is that computers are getting twice as fast every couple of years, while costs stay the same. The theory is based on Gordon Moore’s observations in the 1960s about the rate at which engineers were able to increase the number of transistors on silicon chips. Although the theory has been applied far beyond Moore’s original observations, many have noted that the curve also describes rates of improvement before Moore’s work, suggesting it is not just a self-fulfilling prophecy. I think this curve is an inevitable function of the structure of physics. We could also describe it as the speed of learning, that as we make things over and over again, we are able to make them better. And generally we do see that this rate of increasing computation power at sustained prices has been in effect for about 60 years now. Although there are limits to digital bits, we see this trend of doubling across various technologies, for example in storage and bandwidth, so the curve appears to be independent of the medium. Looking ahead, I expect we will continue to see this curve apply to new frontiers of computing: quantum, photonic, and analog. This curve is so foundational to our experience and expectations of technology, that it defines our understanding of modern life and progress. I feel confident that development in these new frontiers will sustain this curve of better and cheaper.

This video on “The Future of Moore's Law” was commissioned by China Mobile as part of an online course. It is one of 36 lecture videos. A version with Chinese subtitles is available at Citic Migu: http://citic.cmread.com/zxHtml/listen...

A transcript of the lecture in English is available here: https://drive.google.com/file/d/1qlAo...