Why is Moore's Law no longer valid?

Valid for over 120 years: a new look at Moore’s law

In one of my very first blog posts I wrote about Moore's Law, named after Gordon Moore, one of the co-founders of Intel. In 1965 he noted in an article that the number of components in an integrated circuit has increased since its invention in 1958 doubled regularly would have. In other words, the performance of computers doubles every two years. Since Corona at the latest, we have all known what one Doubling over and over again means: yes, it is exponential growth!

Typical exponential growth curve. Here: doubling of the initial value 1 in 38 cycles

The view that Moore’s Law alone on the one that is getting better and better Computing power is widespread. We are all somehow aware that computer chips are increasing their complexity faster and faster at almost constant unit costs. But what if it is just one of many possibilities the reading of Moore's Law? What if it doesn't just apply to chip manufacturers like Intel?

In his book, first published in 1999, “The Age of Spiritual Machines" [in German: "The intelligence of evolution"] Shows the American author, inventor, futurist and Director of Engineering at Google, Ray Kurzweil, the Calculations per second on that one can buy for a dollar. Adjusted for inflation over a long period of time. And suddenly it becomes clear that the computing power of a smooth, exponential curve follows. Since then there has been data in some kind of adding machine you can draw a straight line on a logarithmic scale: the year on the X-axis and the computing power per dollar on the Y-axis.

The 120-year variant of Moore's Law; an exponential curve, shown here logarithmically

And suddenly Moore's law is no longer transistor-centered. Better still: this abstraction enables one longer term analysis. What Gordon Moore, a pioneer in the early computer industry, observed was nothing more than a snippet of a longer-term trend, the even philosophical Raises questions. This trend was also followed by five fundamental technology changes (from low electromechanical pocket calculators to relays, vacuum tubes, transistors and integrated circuits, see figure above) not interrupted. He has correctly predicted the computing power a dollar can buy for over 120 years.

Every point on this line (it's actually an exponential curve!) Is at the upper end of the computational power of its respective time. In the 1890 United States Census, a Punch cardmachine used; the "Enigma" cipher machine used by the Wehrmacht in World War II was used by the electromechanical so-called "Turing bomb" cracked; and a Vacuum tube computer predicted Dwight D. Eisenhower's victory in the 1956 presidential election. Even if the diagram is updated for the time after the book has been published, one can quickly see that the latest processors and especially graphics chips such as the NVIDIA GPU are also on exactly this progress curve are - until today.

Detail of an Enigma encryption machine

But away from the technical details and back to the line: every point, every machine on it stands for Milestones in Progress and often for human dramas. Before Moore’s Law, which, as mentioned above, was first formulated in 1965, probably none of the people on the curve knew that they were actually on one. However, this pace of innovation is working for over 120 years now externally influences the economy and promotes it in the process. Think about what happened during this time: the Great Depression, the First World War, the Second World War and several recessions - none of this has slowed the long-term development of Moore's Law. Certainly there have been fluctuations in the rates of adjustment, revenues, profits and economic fortunes of the computer companies hidden behind the various points. But it's not about them, it's about the long-term trends. And these trends are clear and clear.

Now we are in one Time of constant change, the disruption. In the technology industry, it is difficult to accurately predict a five-year trend, let alone a trend that spans centuries. So could it be that the line from Ray Kurzweil's book is actually one of the most important graphs ever designed?

After all, a large - and growing - number of industries depend on sustained exponential declines in the cost of computing and storage capacity. Moore’s Law drives manufacturers of electronics, communications and computers in front of it. Information technology is increasing too the main driver in the fields of Drug drug research, Biotechnology, bioinformatics, medical imaging and diagnostics. As soon as Moore's law exceeds critical limit values, a laboratory science that was previously based on trial and error can turn into a Science of simulation calculation walk. And the pace of precisely this progress accelerates dramatically - or exponentially ...

Information technology is the biggest driver of the world economy

An example to clarify: Airbus builds airplanes. In the past, the construction services were based on model airplanes in Wind tunnels tested. Since computers have become powerful enough to precisely simulate fluid mechanics, new types of aircraft can move at a rapid pace from iterative simulations to be developed. Wind tunnels lie fallow or are being rededicated for body flying and other fun sports. The last type of aircraft whose design was still being tested in the wind tunnel was the Boeing 777. Today's aerospace engineer simply sits at his desk - and not just when he is working from home. It will be similar with the Research on new drugs. Here, too, a shift is imminent. Less and less will take place in laboratories and clinical tests and ever greater proportions in advance can be simulated by high-performance computers, supported by artificial intelligence.

Sooner or later, every industry will become a data and information business. Jevery branch. What do you think a farmer will tell you 20 years from now if you ask him how to stay competitive? Of course, it will mostly depend on how he is information exploited from satellite images. He uses it to manage his field, which was tilled by machines and robots, and to optimize the code. Namely the Program code in its seeds: the genetic code. Manual processing or even handwork will only make up a small part of his work. It will establish itself similarly in every industry. Information technology pervades the global economy, turning it into a kind of neural network.Let's get to the philosophical part: “Why?” Why has this trend continued for 120 years? Obviously he did nothing to do with the semiconductor industry. However, our current pace inevitably means that we are constantly developing. We are constantly making progress. How can that be?

All new technologies are combinations of existing technologies, Recombinations of previous ideas so to speak. Innovation never takes place in a vacuum. However, if you can combine ideas from earlier times in almost any way you want, it is a bit like - as it is called in the English-speaking world - on the shoulders of giants stand. Regardless of the academic field, today's advances are always built on a great foundation of past experience and inventions. In this way it can also be explained that important innovations often at the same time are ripe and can then be discovered by several people almost at the same time. Just think of Louis Pasteur and Robert Koch, who in the second half of the 19th century described microbes as the causative agents of dangerous diseases within a short period of time.

Progress is based on the correct composition the right ideas. Before the age of science, people mostly relied on chance discoveries. Science became the Processes for innovation but fundamentally optimized. The greatest progress in human history was only made possible through scientific methodology: the targeted collection of knowledge, the well-founded comparison of results, falsification, verification and all that regardless of the political mood - economic growth and technological change are based on this conceptual basis to this day .

Progress is like a big gearing - driven by our ideas, our curiosity and our knowledge

Add to that now new ideas and new thoughts In addition, there are more and more possible combinations and results. The probability of success increases. This explains the innovative strength of our modern way of life through the achievements of a networked and globalized world. It also explains why interdisciplinary ideas are so much more disruptive than those that only come from within an industry. Mixed disciplinary teams are the best starting point for disruptive and innovative ideas Walls within academic disciplines tear down. And the more disciplines involved, the higher the number of possible innovation combinations.

In the last two decades we have also learned to understand the effect of another new factor: the Internet. People can exchange ideas today like never before in human history. Categorizing and searching, similar to the scientific criteria already mentioned, is included. In the past, the success and failure of entire nations depended on the innovations of their own society - today it will together worldwide researched a vaccine against Covid-19.

The internet connects the whole world

Due to demographic change, in the next four to five years for the first time three billion new heads take their first steps on the internet. You can use cheap smartphones and satellite technology from companies like Starlink online even in developing countries go and join this global conversation. Today these people are just “out there” doing agriculture. You don't communicate with the rest of the world. They don't bring in the ideas with which they could potentially enrich the global conversation. This is a rapid influx of three billion people and their thoughts and ideas into the world economy unprecedented in human history. That means one thing for the pace of progress and innovation renewed burst of non-drying sources.

And for Moore’s Law that means that - regardless of the size of the computer chips - a Flatten its curve is far from in sight.

I like it:

LikeLoading ...