Updating hardware and software

by Roger Bourke White Jr., copyright July 2017


We are experiencing an ongoing crisis in updating hardware and software: How to keep the hardware and software behind all this Big Processing and Big Data up to date?

We have this crisis already in the 2010's and it will still be with us in the 2050's: Hardware and software change so quickly that 20 year old equipment and programs feel ancient. But in many instances they are still used. A high-profile 2010's example of this happening is in US air traffic control and airline scheduling systems.

This is going to stay a problem, but solving it is something that can be included in structure design, if it is thought about ahead of time. Basically, it means adding redundancy to the design so that parts of the system can be shut down and outright removed without disrupting the whole operation. Redundancy is expensive, it adds cost, essentially doubles it, but it is another form of insurance. This means it looks expensive until the time comes when it is helping to save the system from catastrophic and anachronistic circumstances.

How self-aware AI systems -- post-"Singularity" systems -- will handle this will be interesting. It will be particularly interesting if these AI systems develop complacent thinking, as in, "I'm thinking just fine! Why do I need an upgrade?" Keep in mind that throughout history complacent thinking has been the enemy of progress and friend of decline. There is no reason to believe this trend will end in the self-aware AI realm.



--The End--