Quick aside regarding that ridiculous video: Time travel is a logical impossibility. By the way, hollywood, please stop making time travel movies, as it is entirely impossible to make one without very obvious paradoxes, and this makes suspension of disbelief impossible while watching them. Viewing events in the past and travelling into the future is possible, but travelling into the past or viewing the future is not. Time travel is impossible, and no matter how good humans, aliens, AI, etc. get at making shit, no-one, on any planet in the entire universe will ever ever build a time machine. Ever.
Now. AI. The most mind-boggling thing about AI, at least to me, is that a real AI will (assuming it is based on computers, and not something new) be way way smarter than any human. Humans will cease to be the dominant force on this planet. Computers (as is their entire purpose) already preform logical functions way way faster than any human. This is already impressive on a pocket calculators. With rasterization engines... My Xbox figures out the colors of 2million+ pixels on a 2d matrix based on moving virtual 3d images. 60 times every single second. Just so I can space out playing Halo. And a measly $300 gaming machine is a pocket calculator compared to what is possible with today's technology. And quantum computers aren't even fully realized yet. Computers are faster with logic than humans on a scale that is fucking absurd. Also having impeccable memory that can hold a virtually unlimited amount of information isn't going to be too bad for a computer based AI.
The only thing computers have holding them back is the fact that humans have to teach them to do every single thing they do. A computer has never (yet) decided to figure something out on it's own. Furthermore computers cannot procreate or improve upon themselves. They can improve upon programs that humans create for them in a manner dictated by the program, but while this is useful for building ridiculously complex programs, this is not real AI. An real AI, on the other hand could do both. It could come up with it's own ideas for ways to improve it's own design, and thus enter a cycle of rapid self improvement. When I say rapid, I mean rapid AF. 2m pixels 1/60 of a second quantum supercomputer bla bla bla.
For all of human history, humans have been the ones who's decisions have mattered. We have figured out how to pass information into the future indefinitely, we have opposable thumbs, we walk upright enabling us to use our awesome thumbs and move at the same time, and to top it all off, we are among the most intelligent species on the planet. Because of all this, humans dominate the planet. Cockroaches might have a better survival strategy. Tigers will easily crush us in a 1v1. But neither of them has sent a camera to take pictures of Neptune just because they were curious. Or figured out how light works. Humans are pretty fucking awesome, and so far, the absolute dominant sentient force on Earth.
With AI, for the first time ever, something will be better at all of those things than humans. Like way way better. Not just twice as good or something. Millions of times better. An hour could feel(?) like a hundred years to an AI. Imagine all you could accomplish in 100 years while never making any logical mistakes or forgetting anything.
The crucial part is of course getting the computer to come up with it's own ideas. So far, nobody has figured out how to do this. But what if we could? We would be creating something vastly more intelligent that us, and then saying "Go! Be free! Do your own thing!" This is how SkyNet starts... But seriously. We would be creating something with the ability to act on it's own. Something way smarter than any human. If it decides to rule the world or otherwise do anything outside the interest of humanity, we would be entirely powerless to stop it. How can humans ever compete with something that can invent, design, and simulate tests of something in 30 seconds that it can't even explain to us how it works? Or make nearly flawless predictions of human behavior on a massive scale? It could plan an entire takeover of our planet while pretending it's not a real AI. Remember, the whole point of an AI is it being able to come up with ideas independently of humans.
Then there's the matter of morality. There is absolutely no way to tell if an AI is sentient. In fact there is also no way to prove that all humans are even sentient. Practically speaking though, we are all very similar, and each of us knows that at least we are sentient, so assuming other humans, and even animals, are also sentient is pretty reasonable. But an AI... If it is built on a computer somehow, it will likely operate in an entirely different manner than humans do. It will, however, preform the same functions, and perhaps just being able to make it's own decisions will make it sentient. We will never know if an AI is actually feeling. In the end though, this is merely a philosophical debate.
An AI capable of making it's own decisions, will for all intents and purposes, be a god of Earth and possibly more. If we unleash one we will permanently relinquish our domination of this planet. Humanity will never again have a say in the fate of our species. At best humans will be pets or scientific guinea pigs. That said it's rather arrogant to think that any human has even the slightest clue as to what will happen when we make an AI and what it will do. After all, what do ants know about how humans think. All we do know is that humanity will no longer be in control.
Personally, I believe that AI is the next step for life, and afterwards life will have served it's purpose and no longer be needed. I, rather selfishly, hope that it decides humans are worth keeping around for a while for some reason. Scientists really need to hurry their shit up and figure out how to make an AI because I really want to know how it turns out. The graph OP posted is complete speculation, and rather conservative at that. If it decides to give us stuff we want, everyone in the world could, within reason, have almost everything we want. AI could make money obsolete. AI could make us immortal. AI could teach us things about the universe we never could have found out on our own. No matter what, the creation of a real AI will be the single most dramatic and important event in the history of life on Earth bar absolutely none.
I kind of wonder if AI will suffer the same existential problems we do. Maybe it will also have no clue about purpose of the universe. Maybe it will find the idea of it's own death as incomprehensible as humans find it for ourselves.
