The influence of Moore's law on the modern world can be more than any other phenomenon. Over the past 40 years, computing power has increased exponentially, changing our lives in a variety of ways, creating a variety of miracles from the World Wide Web, smartphones to the Internet of Things.

But even if Moore's law has not failed, its effectiveness is rapidly decreasing. The big question that has not yet been answered is, to what extent is it important?

In a 1965 article, Gordon Moore, the future founder of Intel, explained that the number of components in integrated circuits doubled every year. In 1975, he revised the cycle to two years, and his prediction (later recognized as a "law") has been basically valid since then.

The impact of exponential changes is often difficult to understand. In practical sense, the improvement of computing performance makes the chip smaller, faster and cheaper, reaching an astonishing degree. The Apollo guidance computer that sent Neil Armstrong to the moon contained 123,000 transistors, while Apple's iPhone 7 had 3.3 billion transistors.

But the natural law of physics is implementing its revenge as it always tends to do. Nowadays, as manufacturers try their best to develop the manufacturing process of 7 nano chips, it seems that we have reached the physical limit of how many transistors can be reliably accommodated in an integrated circuit. The relaxed days that chip designers have enjoyed over the past few decades are at an end.

This seems to be a problem for the computer industry -- and for the rest of us -- because future performance improvements will involve much greater costs and complexity. But chip designers seem to enjoy this challenge. Some even predict that this will open an exciting new era of innovation.

Mike Muller, chief technology officer of ARM, a chip design company, said their products would continue to improve in the foreseeable future, although computer users were unlikely to notice much difference. "Some people say Moore's law is slowing down, and Moore's law has disappeared. But from the perspective of ARM, we don't care, "he said. "The final product will become better and better even if our work becomes more and more difficult."

Inevitably, chip designers have to become more creative. They can "cheat", improve the infrastructure, or stack the chip into a three-dimensional structure. They are also producing chips for special purposes, such as artificial intelligence.

For example, John Hennessy, chairman of Alphabet, said that Google's Tensor Processing Unit, a hardware chip designed for machine learning, can perform 92 trillion operations per second, about 100 times the throughput of general purpose processors. He said at last month's Zeitgeist Minds conference: "We will make progress in this way when Moore's law is not as easy to deliver as it used to be."

The wider changes in the industry have also expanded the way users use powerful computing power. Jason Zander, executive vice president of Microsoft Azure, says computing is increasingly moving up to the "smart cloud" and out to the "smart edge". Nowadays, it is easy to get the powerful computing power of Intelligent Cloud at reasonable cost. At the same time, it is more and more likely to process data on the edge (all kinds of network devices) locally. "We are in a magical turning point in the process of computing history," he said.

Nevertheless, the race to develop new computing methods is unfolding, which will enable us to transcend the limitations of the Silicon Age. The most interesting thing is quantum computing, which is expected to greatly improve its performance, even though its development is still very difficult.

As the United States and China compete for the leading position, the arduous efforts in this regard are gaining a geostrategic dimension. In July, the Defense Department's Advanced Research Projects Agency (DARPA) announced a five-year, $1.5 billion plan to encourage breakthroughs in chip architecture, materials and design. This is part of the electronic revival Initiative (ERI) of the Bureau.

At the launching ceremony of ERI strategic project in San Francisco, Kristen Baldwin, Deputy Assistant Secretary of systems engineering at the U.S. Department of Defense, made it clear that the United States was launching a computing competition with China. "We want to coordinate our common needs to counter China's desire to be the leader of the next generation of semiconductor products," she said.

In addition to fierce commercial competition, chip design is becoming an urgent issue of strategic interest to the United States, which may be good news for all computer users. Historically, competition among superpowers is an excellent way to promote innovation.