Autonomous agents

For what I know, the programs that do interact with the real world are drivers, that modulate the signals to / from hardware. They are at the interface of physical and digital.

On the other hand, the concept of autonomy implies self-determinism, as described in the article on autobots.

Thus, what you are asking to think us about, is the self-determining drivers: imagine a Mars rover, that decides by itself, if it wants or it doesn't want to perform its interface function between the Mars and us, and determines itself how it would like to modulate the interface between us and its hardware. For example, if it gets cold, it may not want to let us move its wheels, or smth like that.

I think, putting autonomy at the driver levels, would mean that drivers themselves could decide to trickle into or infect a computer, and make it work the way they modulate. Biological analogy: having bacteria (autonomous agents) inhabit guts (the interface between digested food and living system).

What could self-determined drivers do for us?

My first thought -- they could wreck havoc to our computer systems, making them unpredictable and life-like. I wonder, what you had in mind while writing this interest :)

I'm thinking in line of Huginn and node red.

https://github.com/huginn/huginn https://nodered.org/

I want some while loop keeping alive my backups and files on multiple clouds. I want self arranging computing - plug in a hard drive and it gets added to the cluster . I'll make ideas for each.