In 1965, Gordon Moore was asked to give a prediction of the next 10 years for the semiconductor industry. In an article published in Electronics magazine, Moore speculated that the number of circuit components in a dense integrated circuit would continue to double every year (this was revised by Moore in 1975 to approximately two years). This became known as Moore’s Law, something of a self-fulfilling prophecy at the centre of the technological revolution since the latter half of the twentieth century.
There are many assumptions and much confusion surrounding Moore’s Law, and it has inspired more observations of the future in technology. Dennard scaling represents performance against power, and this approximately correlates with Moore’s Law. Other factors associated with circuit components (size, cost and speed for instance) were later linked to Moore’s Law or given their own ‘laws’. Moore’s Law is a misnomer too; it isn’t a physical law, but more so an observation which has proven true since its conception.
While there are discrepancies regarding Moore’s Law, it is widely accepted that it can only remain valid for a finite time.
While there are discrepancies regarding Moore’s Law, it is widely accepted that it can only remain valid for a finite time. In recent years, there have been many predictions of when Moore’s will become irrelevant. One of the most vocal on the subject is Gordon Moore himself, who has re-evaluated his estimate several times over the last twenty years. His most recent prediction places the demise of his Law between 2020 and 2025. This is especially significant because in the last decade, many scientists have also placed the end of Moore’s Law in the same approximate timeframe. The most generous, but definitive estimation of the Law’s lifespan is from a report co-authored by Glenn Starkman of CERN, which found that it can only exist in a technologically advanced society for 600 years at the most.
The end of Moore’s Law is significant in more ways than one. It marks a huge milestone toward artificial superintelligence, the creation of which is believed to kickstart ‘the singularity’, causing growth at a level which is incomprehensible to us at present. Basically put, when technology and computers reach an operating power beyond that which the human brain is capable of, there will be a rapid acceleration, an explosion of technology and intelligence. The significance of Moore’s Law here is that it marks a path toward this singularity, paving the way from electromechanical computers over 100 years ago to the nanoscale operations today and beyond.
The indicators that Moore’s Law is slowing grows as time progresses. The estimate of Dennard scaling has already ceased being relevant since 2007. The physical capabilities of the circuit are also a determining factor; as the size of the circuit moves to a scale beyond which the flow of current would be impossible to maintain. A transistor was developed in MIT last year which can carry a single atom across 10 nanometres. 3D transistors are in production which increases the computing capacity of the transistor. Of course, there are ways around this, so it doesn’t have to spell the end of Moore’s Law, but this is merely postponing the inevitable.
Intel, of which Moore is a co-founder, has always taken pride in its ability to maintain Moore’s Law, but even they are aware of the difficulty it has become to maintain the exponential nature of the law. The output has slowed from every two years to two and a half years, and the architecture is not shrinking at the dramatic rate at which it once had. This is not necessarily a bad thing. Until relatively recently, it was unthinkable to consider the once large-scale operations that companies, such as Intel, now perform on a miniscule scale. The progress with this technology may be slowing, but this is entirely understandable given the focus on it over the last century. This is an opportunity to adopt a different approach, a different outlook for moving forward.
Modern computing had reached a point where complacency became rife, and the industry became focused on selling updated shrinking iterations of existing technology. The best example of this is in smartphone technology, where smaller eventually stopped meaning better. The architecture then changed, with attention turning to new features such as curved screens, voice control and near field communication. We can expect this same shift in focus when Moore’s Law reaches its peak, where investments in downsizing will stop and instead research will grow in other areas.
The best example of this is in smartphone technology, where smaller eventually stopped meaning better. The architecture then changed, with attention turning to new features such as curved screens, voice control and near field communication.
One of the biggest disadvantages of this exponential growth is the obsolescence that accompanies it. While the aim of the development is to lower the cost of the technology by making it more accessible, it has become costly to constantly remain up to date with the most current technology. If the technology at the current size does become widespread due to delayed obsolescence, it would give the general public a chance to catch up with the most current iterations.
While the end of Moore’s Law looms closer, even if we don’t know exactly when, we can acknowledge the achievements and advances it has brought to our lives. It is older than the first mobile phone, the Internet, and many products of the technological revolution which it played a part in conceiving. The exponential nature of Moore’s Law correlates with the exponential nature of technology, and it is an exciting future that Moore helped shape.