Your brain really is just a computer. Neurons receive inputs (via synaptic membranes) and depending on different factors the neuron will either fire or it won't. That's binary, a 1 or a 0, just like the transistors that your computer uses. Under the hood it's really nothing a computer can't do or simulate.People have always been stupid whether you tell them or not. When you bragged about how Google was not stupid enough to do what Apple is rumored to do, you yourself opened the discussion to the topic of stupidity.
I know my neural networks well enough and I'm not very impressed with them or any other AI topic. Research in artificial intelligence has always captured peoples imagination, but never lived up to the dreams. Meanwhile powerful handheld computers became real in the past eight years, because of what Apple started with the Newton and finished with the iPhone. ARM chip design and mobile OS design are boring beyond compare, but they solve problems and enhance peoples lives like nothing during my lifetime. Mobile is a story bigger than the Internet. Refugees use smartphones on their flight, neural networks not so much.
I feel confident to proclaim that real intelligence in artificial form does not exist and likely will never exist. Once you know how a chess computer works, it becomes way less impressive. It's all just deterministic programming. Same input, same output. AI software only mirrors the amount of intelligence the programmer put in it, it doesn't come up with a single original thought of it's own.
What else? My iPad has three million transistors and will be outdated by March 15.
You are infinitely far away from the intelligence of a three year old. Only in an environment, where everything is set and well defined AI can find a pre-programmed solution. This solution can be better than that of any human being. A chess computer can beat a grandmaster in a tournament, but when you ask the same program to play a game of solitaire, tetris or bomberman it does not compute. That is the opposite of intelligence.
That's just computing with massive amounts of data. Google does it every day. Intelligence is not about to find the best possible solution, intelligence is to find your way when there is no known path of how to come up with a solution to begin with. When I take your keys away, you will come though the door no matter what. Call your mom, grab an axe, climb through a window. You will find a way once you've identified the problem, because you have intelligence.
Yeah, we call them computers and they are not intelligent themselves. They are tools we humans use to execute our own ideas. Only a few dumb components to create a skateboard and what breathtakingly complex stunts a skateboarder can do with it.
Self-driving cars are easy once you've reduced the task to require no intelligence at all. Here's a list of self-driving subways. Should driving on streets require to find solutions to unforeseen problems, computers will never make it.
Jesus, Google needs to find a way to stop Samsung from forking Android and taking it into another direction, like Amazon did with Fire OS. Do you really think Google Now is the answer to the companies competitive problems?
Speech recognition even lags the intelligence to understand what I'm saying let alone what I mean. Human communication requires the context of a whole human life, history, community. Every single word carries dozens of ideas and concepts thought about since centuries. When I start to speak German, you won't understand a word and you are an intelligent human being. Siri times 100 is still stupid as ****. Any progress just shows how far away AI still is from accomplishing anything actually intelligent.
The first true 'hard' AI are not likely to be created directly by software routines. They're more likely to be enormous neurological simulations that evolve into sophisticated neural networks for doing things like speech recognition. They will likely just use electron microscopes to get detailed 3D maps of the brain, like Broca's area for speech, and simulate it in a computer.
DeepMind recently created a neural network that literally learns how to play most video games (even 3D games), and after a small amount of time can play it better than any human. The number of fields where humans can beat computers grow smaller every single day. We're reaching a time where it's not just clever programming tricks, computers will genuinely be more intelligent than we are, faster, and more knowledgeable. There's just not much left that computers can't do. Some day soon, even consciousness itself will be the domain of computers as well.
The idea that computers will never be able to be sentient is laughable to most legitimate neuroscientists, who realize that the brain is just a large and complicated neural network. The idea that computers cannot be sentient seems to rely on some spiritual belief which has absolutely zero evidence to support it. Not only will computers become sentient - it will happen in the near future too. Advances like quantum computing, etc. will bring it forth faster than most people realize, areas of tech that Google happens to heavily invest in.
THIS is the real future of tech. Research in areas like this are the reason Google/Alphabet is on top today, and it's the reason they'll stay that way long after Apple leadership has destroyed the company with their silly gaudy watches and disastrous cars.
Last edited: