7.8 Future computational technology timeframes

The current wave

We will see continued developments in CPUs, GPUs, and and rapid growth in some early commercial Neuromorphic Computer processors over the coming years as these technologies are well advanced and there are numerous commercial and scientific research applications using these technologies today. 

The first wave of change

Neuromorphic Processing is a particularly active area of research and development, with several research institutions and companies including IBM, Intel, and Qualcomm investing in the development of Neuromorphic chips, in addition to BrainChip currently offering an early stage working commercial products [43]. However, it is still early days for this Neuromorphic Processors, but given that it is heavily focused on neural network computation, it is likely that this will see early rapid adoption in commercially available AI applications.

Quantum computing is the most Computationally advanced of these Computing technologies and has seen significant progress in recent years, with several companies including IBM, Google, NASA, Microsoft and Lockheed Martin developing and/or using early quantum Computers, primarily supplied by DWave. However, quantum Computing is mainly used for helping to solve extremely niche and ordinarily intractable R&D problems, and it may be several years before it becomes broadly commercially available for AI applications. The other issue with Quantum Computing is that it poses serious risks to economic and national security, due to its ability to potentially break all existing commonly used encryption algorithms.

The second wave of change

In relation to potentially more Computationally advanced technologies, the exact timeframes for broad commercial availability is really yet to occur. There are many more significant technical challenges that must be overcome first.

Photonic Computing and spintronics Computing are in the very early stages of development, with researchers focusing on developing the fundamental building blocks of these technologies. It may be several years or even decades before they become commercially available for AI applications, excepting that it is possible that AI could be used to shorten this time frame.

DNA Computing is just an emerging technology that is still in the extremely early stages of development. While it has shown promise for certain types of Computational optimization problems, it is not yet clear whether it will be practical for large-scale AI applications. It may be several decades before DNA Computing becomes commercially viable for AI applications. Again, it is possible that AI could be used to shorten this time frame.

Massive investments required

It is important to note that the development of these leading edge alternative Computing technologies is a long-term process that collectively will require further billions to trillions of dollars investment in future research and development. While it is a near certainty that these technologies, and potentially others, will radically advance the Capabilities of AI, it may be several to many years before they become widely available for commercial AI and AI research by governments, companies, and the general public.

Exponential Learning Intelligence

The most important thing to understand is that AI may be used to accelerate the development of leading edge alternative Computing technologies. This means AI will be able to generate superior Computer information processors, possibly even devising its own exotic technological designs that presently don't exist. 

All Humans then need to do is enable AI with these these superior Computer information processors. This will increase the total information processing rate and Capabilities of AI, which will enable AI to perform more Intelligent operations, such as creating even more superior Computer information processors. AI building AI.

If Humans decide and enable AI to perform closed-loop development of AI by providing AI the ability to build and use its own developed Computer information processors, then AI's growth in Intelligence will almost certainly grow exponentially.

Doing this would be a huge mistake, but unfortunately Humans need to expect that a group of fools somewhere will do it !