Computer 'no longer constrained by limits of human knowledge' Even more startling, the updated version of AlphaGo is entirely self-taught — a major step towards the rise of machines that achieve superhuman abilities “with no human input”. http://www.foxnews.com/tech/2017/10/20/googles-artificial-intelligence-computer-no-longer-constrained-by-limits-human-knowledge.html?cq_ck=1508439127101 [Take a machine and make it as smart as a human being and it will do what humans do so well...think for itself. It will create things never known before, expand knowledge and reshape how we think. I don't believe one can say that this new computer is not as smart as we are. The only major difference is its not sentient. What you think of this?] -------------------------------------------------------------- "libido sciendi"..... the passion to know.
Did you read the entire article? The computer is not actually that smart. It just has better gaming algorithms. Rather than just trying all possible combinations it took a different path to establish the best strategy and tactics. Now that *is* useful but it didn't write its own algorithms. It's just a better "calculating" machine!
No, its just a better calculating machine. From the article, the machine was given the rules and then it basically just played a huge number of combinations (the article even says it "started with completely random play") to find winning moves. The fact it found new moves is not surprising in a game of essentially endless possible move combinations. It didn't "think". Its just an algorithm.
If they can find a way to reward a computer for its behavior maybe some sentience would develop, possibly with the new quantum ones. https://en.wikipedia.org/wiki/Quantum_computing The only reward I can think of is more electricity and additional memory chips.
The "limits of human knowledge" line is WAY over the top. Humans designed a learning algorithm and focused it on a very constrained problem (in terms of board size and rule count). It isn't as if this robot figured out biology, brought history to bear on international relations, or made a decision based on principles of ethics or morality. I do believe this direction will have a significant impact on the future job market, though.
What's missing is a reason to believe either the "creations" or "knowledge" will have any real value. How so? What you're leaving out is sapience, which is at the very least a quantum leap beyond sentience. First you'd have to build a computer that cares about something, and nobody has a clue as to where to begin.
I doubt it will ever be able to create something new, which is not dependent upon known knowledge. It will be able to take what is in its memory banks, and perhaps come up with something that uses the known, to create something that uses the known. But many of our leaps forward into what was not known comes from something else other than what our brains had stored. Tesla got the idea for his generator years earlier, by looking at a sunset. The parts for that generator came to him as it appeared in his mind, while walking, complete, with every part in place. Human creativity can depend upon stored knowledge, which the brain puts together in new way, but it also has comes from somewhere else, where something not in the memory banks, suddenly appear as an epiphany as out of nowhere. I doubt even the most advanced computer will ever do this. It will always be limited by knowledge stored, and will not be able to move outside of that. Yet it will be able to do more than we can, using stored data. But intuition? That might not be possible.
Yes, I don't believe we know how to make a machine conscious, and our applications of robotics tend to be highly focused tasks such as the one in the OP. Surely there will be large value, though. These devices can step in where humans have problems (dangerous environments, ability to stay focused and work huge hours, ability to not be intoxicated, etc.) - factories, roads, policing, etc. Plus, they will be able to bring to bear more information on various problems.
