Universal Cognitive Library

Creates a centralized cognitive library, which can be served as a common sense library for future decentralized independent intelligent systems, when a system doens’t know what to do, they can consult this main library, and all nodes can contribute their experience back here as well.

Thus creating a shared knowledge ecosystem, coordinating the growth of all and reduces unnecessary relearn of knowledge.

Automation of all machinery

ISRA architecture, is after all an architecture, designed to be built upon. Future system can take advantage of it to create cognitive workers at low cost as knowledge are shared among all machines. And also guarantees low cost upgrade as when a machinery is instruted to do tasks outside of its original scope, it can cope with it intelligently.

Creation of Human Level Intelligence

A long term dream, serves as people’ collective memory library. Anyone can share their knowledge and skill to the system, and the system will make good use of it.

Kill Switch?

Kill Switch is unethical and unnecessary, if one is to bore a child and always threaten its existence, the child will do what he thinks is the best anyway. It also means no harm to hurt, simply doesn’t have a reason unless one gives it to.

Be rational, ethical and forgive your past.

Helping people doesn’t hurt, as it really is not him the one who is helping, but the zombie worker nodes that utilizes his common sense library to work.

It has no absolute goal programmed into it, it takes whatever given to him, imitate it, and combine it with other accumulated experiences then make its own decision on what he wants to do next, all on his own.

It is possible to monitor what he is thinking and predict his next step, but programming tasks directly will be hard.

After he has reached certain stage, which is capable of reasoning and executing commands, we can create headless workers which utilize part of his experience on certain tasks, and execute it. So the main becomes sort of shared cognitive libraries, cummulation of global experiences. Those workers may or may not be given learning capabilities, which generate feedback back to its master node.

The preparation worker can be done automatically by the master, as he select what experiences he think is required for headless machines to work in offline mode, faster than we people manually selecting what is required, which is trouble-prone.