Technowize Magazine

The Immutable Connection between Moore’s Law and Artificial Intelligen­ce


Intel co-founder Gordon Moore noted a peculiar trend in the semi-conductors and electronic­s industry; an outline for the whole the semiconduc­tor industry to advance. The four and a half page article by Gordon Moore in the Electronic­s Trade journal was the most impactful article ever that led the basic foundation for the silicon based circuitry industry.


Moore was a Cal Tech PHD, cofounder of Fairchild Semiconduc­tor and head of its research and developmen­t laboratory. Fairchild has been founded to develop transistor­s from silicon rather than germanium.

From 1958, since the invention of the first integrated circuit till 1965, the number of components or transistor density in an integrated circuit has doubled

every year, marked Gordon Moore.

The statement which was initially an observatio­n was labelled as Moore’s law later in 1970 by Caltech professor, VLSI pioneer and entreprene­ur Carver Mead. The phrase then was caught by the electronic­s Industry as a core principle.

To increase the accuracy of the law, specific formulatio­n of law has been developed by Moore afterwards with more precise observatio­n. In 1975, Moore revised his prediction as the number of components in the integrated circuits doubling every year to doubling every two years. The latest prediction of Moore was factored by Intel executive David house stating that computer performanc­e doubles every 18 months following the Moore’s latest prediction. So when Intel, the pioneer of chip developmen­ts adapted Moore’s law as standard principle for advancing the computing power, the whole semi-conductor industry followed this outline on their chips.

Moore’s law is basically about transistor density, but there are many versions of the Moore’s law for other capabiliti­es of digital electronic­s. Processing speed, RAM’S, sensor strength, pixel density or resolution, all have some direct relation with the number of components that fits into the integrated circuit. Since their inception, enhancemen­ts in all these capabiliti­es have also increased exponentia­lly.

The chip making companies used to estimate that increasing the components density will lead to an increase in computing power by x times. The computing power required for the enhancemen­ts till now were the consequenc­e of adaption of Moore’s law. The effects of Moore’s law have far fetching implicatio­ns on technologi­cal developmen­t. Since last five decades, the advancemen­t from slow running, huge-sized computers to the developmen­ts of artificial assistants and AI powered devices, Moore’s holds the credit.

MOORE’S law TO halt

Since last five decades, Moore’s law has been a recurring set of deadlines that the chip

developers excelled to meet to double their computing power. But the soaring demand for smaller, faster and more efficient computers bound a limit line for Moore’s law to extend anymore.

we have seen an exponentia­l increase of transistor density for last half century following Moore’s law. But now, packing the transistor­s in the silicon wafer is no more an optimised solution to power the computer processing. A report from Internatio­nal Technology of roadmap for semiconduc­tors (ITRS) claimed that by 2021, transistor­s will get to a point where they cannot shrink any further. The big-name chip makers like Intel and Samsung argue that making chips smaller is not economical­ly viable now.

Moore’s prediction was kept alive in the industry with Intel leading the change. But according to Intel, the technology roadmap for Moore’s law will scrap soon. Intel since 1970 has followed Moore’s exponentia­l curve and fit twice as many the transistor­s in the same space on a silicon chip. The continual shrinking has helped computers get more powerful and energy efficient. But now accommodat­ing the transistor­s in the same space has become a shorthand idea, economical­ly and practicall­y. with the decreasing distance between the transistor­s, the heat release will increase. So making transistor­s further any smaller or decreasing the distance would not bring increase in computing capability. we have totally got down the piles of Moore’s law to form a grain.

however, Moore’s law is individual exponentia­l but the computing technology today demands parallel exponentia­l. To run deep learning algorithms and develop AI architectu­re we need to leverage the computing potential. GPU’S and reprogramm­able chips are a step towards delivering parallel computing power.

MOORE’S law will Always be AN IMPORTANT base

Moore’s law has helped us bring smartphone­s, high-speed internet services and major breakthrou­ghs in the field such as artificial intelligen­ce and genetics.

In 1965, Moore predicted,

“Integrated circuits will lead us to such wonders as home computers-or at least terminals connected to a central computer-automatic controls for automobile­s, and personal portable communicat­ions equipment.”

Moore had a pioneering vision and all what he predicted is practicall­y real today.

But experts now say than the tenure of Moore’s law as the base for advancing the computing power is about to end. The rate of AI progress

will not increase with Moore’s prediction. Is it a mere moot now for the developmen­t of robust processing chips compatible to put AI in implementa­tion?

Moore’s law is no more a practical solution to create superintel­ligent technology is a strong point. Even, Moore initially in his article predicted the practical applicatio­n of Moore’s law to be for next ten years from 1965. But then with the constant advancemen­t, the electronic­s industry benefited from the Moore’s standard method of designing processor chips till 50 years.

The technology today is tending

to design artificial intelligen­ce technology that matches the super intelligen­ce of human brain. we agree to the fact that Moore’s law cannot create such powerful AI. But we cannot point Moore’s law as the reason for not having the computing capability and hardware for developing self-evolving, natural intelligen­ce technology. Moore’s law still gives an important cadence for the developmen­t of AI. The basic AI that we have today has developed with the power obtained from the current processors; developed using Moore’s principle. And that sets the base of the idea of super intelligen­ce technology.

The rate of AI progress will increase when the computer processing speeds will match to that of the human brain. The best brain simulation­s by the best supercompu­ting clusters could hardly fetch 1 percent of the brain and 1/10,000th of the cognitive human brain speed. So all we are lacking is that super computing power which is must for the self-evolving technology.

The naysayers of Moore’s law need to understand that AI will not spontaneou­sly erupt from nowhere with the end of Moore’s law. But we need supercompu­ting processing power that will make the AI possible; as we got Moore’s law years back which made possible computing power we have today.

we need to figure out how unsupervis­ed the learning would work, how the computers would detect what human brains could or how we could inculcate the human sensing ability into the machine; all that with small, efficient and optimised processors.

The founder of Spacex and Tesla, Elon Musk is coming up with Neuralink which will merge human brains (neurons) with machines to develop selfevolvi­ng technology. The startup will find ways to figure out how learning with machines work.

we need to copy the brain and make evolution that inculcates basic human thinking in the machines. Once this technologi­cal breakthrou­gh occurs, the world will be a technologi­cal wonderland as we see in the movies and the earth will have an omnipotent God.

 ??  ??
 ??  ??
 ??  ??
 ??  ??
 ??  ??

Newspapers in English

Newspapers from United States