Been there, done that.
In fact, my thesis was about using neuroevolution (using evolutionary algorithms to evolve artificial neural networks) to solve a specific problem.
In terms of organic evolution, I think it is based purely on chance. If one gene mutated differently in the evolution of humans, we could look completely different to what we do now.
It is about chance, but it is also about statistical averages. If one gene mutated differently for one person only, it would have to be pretty damn beneficial mutation for it to spread through the whole population.
Also, mutation only survives long times if it is beneficial. And whether it is beneficial or not depends a lot of the environment the individual is living in. So while it is true that we could look very different, there would probably still be certain similar qualities simply because they work in this kind of environment.
A million years is short for evolution... but that doesn't mean it cant happen in a million years. There could be some super organism that evolves 10 times as fast as we do. Scientist base their views on the speed of evolution on what we know now. It is unpredictable for the future.
You don't even need fast evolution for that, necessarily. Since evolution is a random process, it is possible (just very, very unlikely) that all the necessary mutations could occur in a single generation.
Although I'm not so sure it is possible with actual biological evolution, since there is the whole embryogeny thing and there are probably some limitations on how much a genome can change in a single generation and still survive. I don't know really, my expertise in on the artifficial evolution side.
Anyway, this is more of a philosophical topic, since the probability of that happening is so low that it is practically zero anyway.
There's no evidence that computers are always right. After all, it is merely a creation made by humans.
You're right, of course.
In fact, there's evidence of computers sometimes not being right (hardware failures and all that; Intel Pentium bug is probably the most legendary one, although that one was a manufacturing error). And there's even more evidence about computer programs not being right, in fact many of them tend to contain some bugs.
However, I'm not even sure one needs a computer for potassium-argon dating. And I'm not trusting computers as much as I'm trusting a group of researchers who used those and other tools to come up with the date for the fossil.
For human standards it is.
Maybe, but universe doesn't follow human standards and in my opinion it would be arrogant to assume it does.
Also, there's no evidence of a living creature
that old.
Yes, I would say it is highly unlikely that any lifeform based on biology as we know it would be able to survive that long. However, I fail to see the relevance of this. I thought we were discussing whether there's any evidence for mankind being more than 6000 years old, for which there clearly is some scientific evidence. Whether you think the evidence is strong enough to be believed, is, of course, whole another thing.
There can be no strong proof about any information we gather with our senses, since senses provably make mistakes and can be mislead by illusions and such. So there's no strong proof about pretty much anything and it all comes down to how much evidence one needs to believe the issue. However, most people tend to simply take the theory with maximum likelyhood (under the currently known evidence) and believe that. Some others choose to believe something based on their personal conviction.
There's been some attempts at making 'independent' machines.
One of the biggest challenges there is modelling motivation, I feel. It is very hard to come up with an artificial motivation function that would allow the independent agents to keep learning but would steer them clear of areas and things where they are no longer learning. So it is very hard to create an independent agent that would be capable of learning and improving itself enough to be able to operate in an environment without any human guidance or help. Especially if the agent is meant to have "free will" and explore things that interest it instead of things that humans have told it to explore.