The future of computing
For the longest time, advances in computing have tracked Moore’s Law. However, as we nudge up against the practical limits of Moore’s Law, the future of computing could lie in a new paradigm that was previously thought to be only theoretical: neuromorphic computing and quantum computing.
In 1955, Gordon Moore, one of the co-founders of Intel observed that the number of transistors on a microchip doubles every two years. Said another way, the ability to more densely pack a chip with transistors has supported exponential growth in computing power. This observation has proven remarkably accurate, although for many years industry observers have predicted the demise of Moore’s Law. Regardless of whether Moore’s Law will continue to remain valid, there are currently advances being made in the realm of quantum computing and neuromorphic computing that could prompt a veritable step-change in what is possible from computing.
Virtually all computers today exist with a layout that follows what is known as the Von Neumann architecture, with a key part of this being data and instructions being stored in binary digits (i.e., 1s and 0s). For a conventional computer the basic unit is a “bit” and it is made of on or off transistors (the 1s and 0s). However, for a quantum computer the basic unit is a quantum bit (colloquially referred to as “qubits”) that have values that can range between 0 and 1, or also 1 and 0 at the same time. What this means is that through a quantum superposition – that is, overlaying the states of 0 and 1 – as well as quantum entanglement (multiple qubits in an intertwined state), quantum computing is able to achieve levels of computational power that are orders of magnitude above what is capable on conventional modern computing architectures.
While quantum supremacy has been claimed, mastering the use of, and the commercialisation of qubits is yet to be achieved (qubit operations are highly susceptible to environment disturbances and are thus prone to error).
Neuromorphic computing, unlike quantum computing, is based on existing chip process technology, but relies on chips that are physically structured like artificial neural networks that mimic the human brain. The beauty of the human brain is its ability to process very complex information in real time, and do so using very little energy. Neuromorphic chips are comprised of many small computing units that correspond to an artificial neuron, with a physical architecture analogous to the synapses in the human brain. The potential of these neuromorphic chips is supported by their fast computing speeds, low power consumption, and, much like a neural system in the human brain, learn, adapt and evolve in real time.
Advances in both neuromorphic computing and quantum computing could unlock potential in analysing much of the data in existence that is still considered “dark.” Dr John Kelly of IBM, estimates that over 90 per cent of the world’s data is “dark,” which refers to data collected by organisations that is not being used; in other words, meaningful insights are not being drawn from this data.
Source: IDC, Morgan Stanley Research
While we have not yet mastered these new technologies, they present exciting opportunities for the ability of firms to process the enormous volume of data they generate, and extract insights. A number of companies are investing heavily in these areas, and it’s likely that over the next 5-10 years we will see developments in the realm of quantum and neuromorphic computing that could see organisations move a step closer to being able to analyse and benefit from dark data.