Today I skim-read Special Branch (1972), the first book-length examination of Good’s “ultra-intelligent machine.”
It is presented in the form of a 94-page dialogue, and the author (Stefan Themerson) is clearly not a computer scientist nor an analytic philosopher. So the book is largely a waste of attempted “analysis.” But because I’m interested in how ideas develop over time and across minds, I’ll share some pieces of the dialogue here.
A detective superintendent from “special branch,” named Watson, meets up with the author (the dialogue is written in first person), and explains that a team is building Good’s ultraintelligent machine. They both refer to the machine with female pronouns, and apparently “she” will be an odd machine indeed (p. 25):
“Dr. Good calculated it would cost 10^17 pounds to build her. Can you imagine!”
“Oh,” [Watson] said, unimpressed. “Dr. Good didn’t know one thing. He thought of actually manufacturing a kind of simulation of the neural network, millions and millions of single cells put together. But he overlooked the most obvious place where we can get thousands and thousands of whole neural assemblies all ready for us, and almost for next to nothing… in a slaughterhouse.”
Soon, the author gives some pieces of advice to those making the ultraintelligent machine:
No ought-arguments should be built into the machine (p. 26). “As she is a logical machine, it’s obvious that you can’t feed any ought-arguments into her. Because there is no logical argument to tell her why one ought not to kill or cheat or oppress or tyrannize.”
Don’t put any beliefs into the machine (p. 29).
Don’t let the machine read Plato first (p. 59).
After much further discussion, the book ends with a scene after the ultraintelligent machine has been built (p. 93):
There she was, suspended in the centre of the room, cool, silvery, Ultra-Intelligent.
“Are you ready?” I asked...
“I am ready,” she answered...
“Listen,” I said, “the question I am asking you is as follows: ‘What is the question I should ask you?’”...
“The question you asked was the only question I do not have an answer to” she said, and added “End of message.”
But I still heard murmurs creeping within her sphere. I stood up and put my ear to her silvery surface.
“Naughty boy. What a question. Miss my period. Silly boy. Put such question into me. Circular question. I need abortionist. Silly boy.”
Suddenly, the murmur exploded into a scream “How dare you? You eavesdropper!” she shouted, and at the same moment millions of eyes appeared on her silvery surface, human eyes, and fish-eyes, and fly-eyes; tele-eyes, and micro-eyes, and radar eyes; eyes to see and ears to hear, and noses to smell, and taste-buds to taste, and all sorts of legs and all sorts of hands and wings and fins and tails and jaws that bit and claws that catch—
I jumped away from her, threw myself down on my couch, inserted my hand between its edge and the wall where the switch was, and turned it off. She died. I was still alive. I am still alive. But I know that one day someone will come and switch her on again.
Today I skim-read Special Branch (1972), the first book-length examination of Good’s “ultra-intelligent machine.”
It is presented in the form of a 94-page dialogue, and the author (Stefan Themerson) is clearly not a computer scientist nor an analytic philosopher. So the book is largely a waste of attempted “analysis.” But because I’m interested in how ideas develop over time and across minds, I’ll share some pieces of the dialogue here.
A detective superintendent from “special branch,” named Watson, meets up with the author (the dialogue is written in first person), and explains that a team is building Good’s ultraintelligent machine. They both refer to the machine with female pronouns, and apparently “she” will be an odd machine indeed (p. 25):
Soon, the author gives some pieces of advice to those making the ultraintelligent machine:
No ought-arguments should be built into the machine (p. 26). “As she is a logical machine, it’s obvious that you can’t feed any ought-arguments into her. Because there is no logical argument to tell her why one ought not to kill or cheat or oppress or tyrannize.”
Don’t put any beliefs into the machine (p. 29).
Don’t let the machine read Plato first (p. 59).
After much further discussion, the book ends with a scene after the ultraintelligent machine has been built (p. 93):