Life After Silicon:
The Age of the Cup-holder
By Shlomo Maital
Most of us are blithely unaware of how much we owe to a law – Moore’s Law, named after Intel co-founder Gordon Moore, which says in simple language that every 12 months or so, the number of transistors in a given area doubles. In a 1965 paper, Moore used 4 or 5 data points to show that this had happened up to 1965, and he said it would continue for the foreseeable future.
And it did..and how! We now carry more computer power in a chip barely 3 sq. mm. than the 1976 room-size supercomputer designed and built by a genius, Seymour Cray.
But what if Moore’s Law is about to be repealed? What if the future ends? As Francis Fukuyama wrote, in 1992, we are at “the end of history”; are we at the end of the history of silicon?
Yesterday I heard a brilliant lecture on this subject by Prof. Mark Horowitz, on Innovation in a Post-Moore’s Law World. Horowitz is a former Chair of Electrical Engineering at Stanford and is visiting my university, Technion. He delivered the annual Hershel Rich Lecture.
Here are some of his key points:
* As we move down from IBM’s path-breaking one micron chip in 1974, to today’s 20 nanometer chips, 50 times smaller, we have reached the limit of Moore’s Law, mainly because the heat generated by these tiny powerful microprocessors is very hard to dissipate. Cooling limits further transistor doublings. Mathematically: “computing power” = energy per operation x operations per second. If we want to boost ‘operations per second’, we need to lower ‘energy per operation’..and we’re reaching the limit on that one.
* Result? “In future, success will no longer be about technology development. It will be about finding the right applications of the existing technology”.
In other words: We have loads of computing power. Now we need to find ways to adapt it to create value. The so-called Internet of Things is misleading. There is no “internet of things”. It is in fact a broad collection of devices, and each device needs its own application, its own software, its own microprocessors, to enhance its value.
Horowitz noted that ITRS (International Technology Roadmap for Semiconductors) has for years done roadmap predictions of future technology, based on Moore’s Law. They have now stopped. No more roadmaps. The end of history.
What does this mean for entrepreneurs and innovators? In future, there will be a huge premium for IT engineers who understand the technology, but who also have deep sensitive knowledge of business, customers, preferences and markets. Only this will enable us to have life after silicon. I think every engineering school should have a required course on Startup Entrepreneurship. Silicon will no longer be a rapid-growth industry, it will become like steel and plastics, notes Prof. Horowitz — big, but stable.
Sure, there may be a huge technological breakthrough. But Horowitz notes, it is unlikely, because developing it will take enormous resources, more than governments can afford, and more than private investors are willing to risk.
Prof. Horowitz has a powerful metaphor. In future, he noted, successful products will include many “cup-holders”. A cup-holder is a place to put your coffee or soda, in a minivan or car. German carmakers started the idea. Today some vans have 16 or 17 such cup-holders. They are low-cost, and are highly appreciated, because they show the car-maker is aware of, concerned about, the lifestyle of drivers.
Innovator: Can you use existing computer power, to create small low-cost add-ons (incremental innovation) that create great value? Apparently, this will be our innovation lives – after Moore’s Law expires.