New Approach to Deep Miniaturization: A Way to Overcoming Moore’s Law

Maria K Koleva

doi:10.4236/jmp.2021.123015

Abstract

The matter about some far-going consequences commencing from the replacement of one of the basic principles of the traditional physics that is the principle of locality, with the recently put forward principle of boundedness is considered. It is proven that the thermodynamic theory which is explicitly grounded on the principle of locality, suffers inherent contradiction which roots lay down to the principle of locality. The way to overcome it goes via the replacement of the principle of locality with the recently put forward by the author principle of boundedness. In turn, the latter gives rise not only to a fundamentally novel notion for a number of ideas but also it yields replacement of the proportionality between the software and hardware components with a new non-extensive approach to the matter. It is proven that the famous Moore’s law stands in new reading both in its support and the way to overcome its limitations. Apart from its role for the technical applications, the present considerations stand also as a methodological example how the role of the basics of any theory affects practical rules such as the Moore’s law.

Highlights

One of the greatest challenges to the general theory of artificial intelligence is establishing of general protocols for organization of the match between software and hardware for different types of intelligence
What is more exciting is that such things as laws and rules which hold for one type of intelligence, e.g. Moore’s law for algorithmic intelligence, transforms to new view point, that is non-extensivity and adaptability of relation structure-functionality for the semantic intelligence
The present paper is an example that practical rules such as the famous Moore’s law hold for specific setting only and that the way of overcoming their limitations goes via fundamental changes of those settings

Summary

Introduction

One of the greatest challenges to the general theory of artificial intelligence is establishing of general protocols for organization of the match between software and hardware for different types of intelligence. It is obvious that the notion of the chemical potential is apparently related to the size, the structural and the functional organization of any system The latter justifies the assertion that for each and every system a threshold arises below which the system is not able to operate steadily. Any hardware (those constituting by a single component included) can execute any software (a probabilistic one included) At first glance this matter becomes extremely difficult since the path dependence of the chemical potential opens the door to functional inequality of otherwise structurally identical species. The replacing of the principle of locality with the principle of boundedness renders the notion of the chemical potential to be apparently related to the notion of stability and renders strong sensitivity on the path through which the association/dissociation proceeds This justifies Moore’s law by means of providing another strong argument in its favor. It is demonstrated that this yields to its diversity through non-extensisvity

Non-Extensivity and Diversity of the Relation Structure-Functionality

Conclusions