Machines are systems that convert energy and do work. Machines come in large classes that can be sorted by industry and activity. Transportation systems such as cars, trains, buses, trucks and airplanes. Construction equipment. Farm equipment. Water treatment systems, power generation equipment, oil exploration, drilling and refining. Thousands of machines are necessary in order to execute all the work that is necessary to sustain modern life in an industrialized society.
Man invents machines with the intent of solving a problem, often focused on making labor less difficult. In recent times we have added the idea of making labor more “safe” for humans. The real basis for creation of machinery is the possibility of making a profit. Making cars that cost less, or being able to extract crude oil from the ground and sell the byproducts are examples of industries that require massive amounts of machinery which can be used to make a profit. And because even human safety has a calculable cost, businesses will pay money to control that cost and minimize the risk of injury. Or, at least they should.
Given that man’s machine building nature has been demonstrated for thousands of years, the modern notion of “flexible” machinery or “programmable” control systems is a relatively recent occurrence. The ability to add sophisticated controls to mechanical systems began with the advent of electricity and the relays controls used in car manufacturing. Relays gave way to the “programmable controller” of the early 1970’s, and as computers became more powerful, control system applications followed.
Programmability is the core capability that has been under development for the last few decades with millions of man-hours of effort. If a problem can be modeled mathematically and coded into a binary language, a computer chip can execute that code. Where things get a little tricky is the complexity of task and the sophistication of code capability. It has taken decades to create features like speech, speech recognition, facial recognition or autonomous searching.
Going out onto the Internet and looking up device drivers is a pretty sophisticated “behavior” that looks “smart”. As machines become “smarter” they will merely be following the instructions created by their programmers. So far, none of our machine control systems has come up with it’s own algorithms of control. None have “decided” that the programming was defective and tried to re-program itself. Nor do we expect this to occur.
Difference between human intelligence and that which we call machine intelligence are extremely significant. That which humans invent cannot then, of itself, invent something new. If you throw a baby into a pool of water, it learns to swim. If you throw a computer or a robot into a pool of water, it is destroyed.
that which we invent cannot, then, invent of itself
does the purpose change