Fundamentally the difference between the two is this ; Life needs no Human interference in order to move , replicate and become . Life is not programmed by Humans , machines are .
Oh Lord if dreams were only real I'd feel my hands on that wooden wheel And with all my heart I'd turn her round And tell the boys that we're homeward bound Here's one more day on the Grey Funnel Line I'll pass the time like some machine Until the waters turn to green Then I'll dance on down that walk ashore And sail the Grey Funnel Line no more
That's kind of Musika's point. It's the same as 'How many flapjacks does it take to cover a doghouse?' He's saying the question is insufficiently formed.
I don't see what that has to do with anything I said. river posted, "Life needs no Human interference in order to move , replicate and become," and I replied, "So, we're not too far from building 'living machines'?" If the criteria for life are, "needs no Human interference in order to move , replicate and become," I think my question covers that doghouse.
I'm inclined to define 'life' functionally, such that 'life' is whatever performs a set of functions exemplified by biological life on Earth. (Reproduction, natural selection, metabolism etc.) The importance of thinking of life that way might arise when and if we ever encounter extraterrestrial life. My expectation is that while any life we encounter will perform the various functions of life, simply because we are defining 'life' that way, it might not resemble Earth life very closely in how it manages to do it. There may be no end of ways to accomplish the same sort of result. Machines already move. (And much of Earth life doesn't.) It isn't much of a stretch to imagine self-reproducing machines, machines that can exploit the natural materials around them (metabolism) and manufacture more of themselves (reproduction). I'm not sure what you mean by 'become', but a self-reproducing machine would presumably be subject to natural selection, just like a biological organism. So they would evolve. So it might be very difficult to predict what they will become in the future. Again, I can easily imagine AI's with the ability to program themselves, with the ability to write their own code so to speak. So, bottom line, I don't see any huge distinction-in-kind between life and machines.
There is no reason to presume a self-reproducing machine could evolve unless it were deliberately configured to do so. A requirement of evolution is the mutation of its blueprints and the incorporation of those changes in the next generation.
What's required is heritable variation and natural selection. If a self-reproducing machine's internal 'blueprint' is reproduced over and over and over, from generation to generation, it's possible to imagine small variations in the code creeping in if the copying mechanism isn't 100% perfect. (And nothing in real life is 100% perfect.) We could say the same thing about the DNA code in living genomes.
Biological evolution is a blind process - but machine evolution could be intelligence-based. A forklift could say, "I need to be taller to reach those higher shelves," and it could build itself taller. Then it could say, "I need a wider wheelbase to stay stable at those higher heights," and it could build in that modification. Then it could broadcast the blueprints for current and future generations to emulate and expand on.
Good thought. Machine evolution might be a lot more Lamarckian than Neo-Darwinian. (The idea that giraffe necks got longer from generations and generations of stretching to reach delicious leaves.) Self-reproducing machines might detect the need and just build in the necessary mods. If so, I'd expect them to evolve a lot faster than Earth life does.
