[QUOTE="Frame_Dragger"]
[QUOTE="Stavrogin_"] I wouldn't know, i haven't played it. :)
Look, i'd hate to sound like a pretentious douche, but can someone please contribute to this discussion, ask a question or make a point instead of a witty remark?
Stavrogin_
I feel for you, so why not. Whatever the movement, the actual distance technologically from such a capability, never mind CHOICE, is vast. I think the issue with many "singularity" models are that they ignore what people WILL do, compared to what they CAN do. The first person who gets their "wetware" hacked is going to start a worldwide movement to abolish the tech, assuming that ever happens. You describe two major challenges, one is creating a meaningful AI (which is frankly as distant as ever), and understanding the human brain so well that we can "link" it effectively.
I have serious doubts that either will occur, and if they do it will be LONG and slow process with many intermediate steps during which I suspect people will find they have issues with brain-computer-brain interaction. Technological singularty theory also generally means an end to humanity as a useful element of the world... I don't see that ending in a term from Deus Ex, just ending humanity.
"The AI does not hate you, nor does it love you, but you are made out of atoms which it can use for something else." (Eliezer Yudkowsky)
Thank you. :D
I believe you're sceptical because of the many scientific predictions that have turned out to be false. Often we hear predictions like this will happen in 10 years, that will happen in 15, and unfortunately, 10 or 15 years after the announcement they say we'll have to wait for another 10-15 years.
The skepticism about AI is mainly due to the overblown statements by the pioneers of the IT industry many years ago, who messianically prophesied that today we would have AI. But the problem here is not only the engineers at that time were unrealistic about the development of hardware and software, but they also had no idea of the complexity of the human brain and the sheer volume of computations it can perform. Today's a different story, scientific teams can replicate a part of the brain in software, there is already a successful brain machine interface based on EEG (so far almost exclusively on motor functions) in existence, but the real ace is in genetic algorithms: Write a bunch of algorithms, leave them to reproduce and make seed, then expose the seed to evolutionary pressure.
That's why i also believe 30 years or so is a realistic number. Technological advancement doesn't have to go way, though. There is a logical but not a necessary link between transhumanism and AI. Integration with nanotechnology is the other way to go, albeit it's a much harder process because if we're being realistic here, the human body is an evolutionary work that is not designed to live more than 30-40 years. The reasons for this are deep and systemic, and without completely rearranging the entire genetic make up, to the point at which rearranged "human" would be far from what we consider to be a human, radical changes are not possible. So it's more likely that we will develop AI that at worst, will just be an electronic equivalent of the human brain. Once this level is reached, this AI can (only) advance with a pace unimaginable for us. Nice eh?
I'm skeptical because of the science I know, which includes knowing that there is nothing that anyone would call a "brain machine". Being able to move a cursor with training is not the same as properly interfacing with the brain, or even interpreting thoughts. Your points seem more aspirational than realistic; nothing in current computer science or human-machine interface even REMOTELY points to a 30 year mark as being close to what you describe. First, you'd need a far better integrated circuit which is going to require a new substrate. Lots of promising materials, and none of them are even CLOSE to mass production.
You need more than the crude ability to hook up an EEG and match the trace when someone thinks, "UP", "DOWN" etc... and then tell a computer to act on that. As for AI, it's as far away as it's ever been; just look at the 'Go' problem. We make computers that better mimic intelligence in VERY restricted circumstances, but it's an illusion that is far from the reality of a thinking machine. It would be realistic to hope that something better than the Monte Carlo system could play 'Go' in 30 years... maybe. It's also possible that the "cocktail part problem" for voice recognition will be handled by then, but the very fact that we have to do so much to get a computer to exist within its narrow range of capcity pretty much kiklls your argument.;
I don't mean this to cause offense, but what I'm reading in your post is a lot of pop-sci concetps that are divorced from the real worl challenges which stand in their way, the current state of the art, and what is required to move that forward. There's nothing wrong with dreaming, but if you want to take it beyond a personal fantasy you have a LOT to learn.
Log in to comment