The whole point of the stories is that it doesn’t work in the end, it is a case study in how not to do it. How it can go wrong. Obviously he didn’t solve the problem. The first digital computer had just been constructed, what would you expect?
Obviously he didn’t solve the problem. The first digital computer had just been constructed, what would you expect?
The FAI problem has nothing to do with digital computers. It’s a math problem. You’d only need digital computers after you’ve solved the problem, to implement the solution.
Not that they weren’t good stories, and not that I expect fiction authors to do their own basic research, but I wouldn’t say they’re about the Friendly AI problem.
It is most certainly not an academic look at the concept, but that doesn’t mean he didn’t play a role in bringing the concept to the public eye. It doesn’t have to be a scientific paper to have a real influence on the idea.
Along those lines, I’d recommend the Metamorphosis of Prime Intellect. It’s a short-novel length expression of an AI that gains control of all matter and energy in the universe while being constrained by Asimov’s Three Laws.
If you want to read a full length Asimov book, my personal recommendation is the End of Eternity. It has a rather unique take on time travel and functions well as a stand alone book. It has just been reprinted after being out of print for too long.
Foundation is his most well known novel and it also very much worth reading.
I can’t find someone violating the copyright online with a quick Google, but Asimov’s short story “The Last Answer” is also a good one with a different take on religion than “The Last Question”.
Ah yes, CronoDAS recommended that, too. (Sorry, I should have acknowledged!)
Oh! More Asimov, “I, Robot”. Here the guy was talking about Friendly AI in 1942.
Not really; they’re not decision theory stories. The Three Laws are adversarial injunctions that hide huge amounts of complexity under short English words like harm. It wouldn’t actually work. It didn’t even work in the story.
The whole point of the stories is that it doesn’t work in the end, it is a case study in how not to do it. How it can go wrong. Obviously he didn’t solve the problem. The first digital computer had just been constructed, what would you expect?
The FAI problem has nothing to do with digital computers. It’s a math problem. You’d only need digital computers after you’ve solved the problem, to implement the solution.
Not that they weren’t good stories, and not that I expect fiction authors to do their own basic research, but I wouldn’t say they’re about the Friendly AI problem.
It is most certainly not an academic look at the concept, but that doesn’t mean he didn’t play a role in bringing the concept to the public eye. It doesn’t have to be a scientific paper to have a real influence on the idea.
Along those lines, I’d recommend the Metamorphosis of Prime Intellect. It’s a short-novel length expression of an AI that gains control of all matter and energy in the universe while being constrained by Asimov’s Three Laws.
It’s available free online under copyright. http://www.kuro5hin.org/prime-intellect/
If you want to read a full length Asimov book, my personal recommendation is the End of Eternity. It has a rather unique take on time travel and functions well as a stand alone book. It has just been reprinted after being out of print for too long.
Foundation is his most well known novel and it also very much worth reading.
I can’t find someone violating the copyright online with a quick Google, but Asimov’s short story “The Last Answer” is also a good one with a different take on religion than “The Last Question”.