How computing is evolving and could enable a better understanding of our economy
Microchips and the broader technology world may soon experience a significant shift. Moore’s law, named after Intel founder Gordon Moore, is slowing, while increasingly complex microprocessing demands are accompanied by the rise of tech trends including machine learning and artificial intelligence (AI). This may be especially relevant in the context of the shift towards a more regenerative economic model, where the notion of systems is essential.
Until recently, Moore’s law guaranteed that computing power would double roughly every two years, while costs continued to go down. The slowing of this process, combined with the pressures created by the emergence of the cloud, AI and machine learning, has significant implications for the microchip industry more broadly, and Intel, its dominant company.
Enormous amounts of data are used by the latest computing technologies requiring more number-crunching power than the world’s largest data centres would have needed only a few years ago. Processors are not improving quickly enough to keep up and the result has been the emergence of a rival to the ‘central processing unit’ (CPU).
Designed to carry out the complex computations required by video games, graphic processing units (GPUs) designed by a company called Nvidia, have hundreds of specialised cores (the brains of the processor) working together, as opposed to a few powerful ones tackling tasks in order. While CPUs have a maximum of 28 cores, Nvidia’s latest GPUs have well over 3000.
Previously used to transform small personal computers into devices that could handle complex games, GPUs are now being used in data centres where AI programmes utilise the high levels of computer power that they generate. In a very short space of time, just about every online giant has now begun to use Nvidia to power their AI services.
GPUs are just one of the specialised processor innovations that are beginning to be applied across the emerging more complex computing tasks.
It is, perhaps, an early consideration at this point, but the capability to handle increasing levels of complexity could provide a significant advantage in the development and modelling of a regenerative economy. At the core of a circular economy, which aims to design better products, services and systems for the 21st century, is the notion that intelligent human-design can take inspiration from and emulate the achievements in nature. Understand and modelling to any level of accuracy has been hitherto near impossible, but the evolution of microprocessing into something more dynamic could be the key to powering the AI and machine learning that supports a better economy in the future.