Friendship is Optimal
is not a "My Little Pony" fanfic, but a SF story that starts with a procedurally-generated MLP MMO, and crescendos to what could very well be the Best Possible Outcome if self-optimizing algorithms are given /almost/ the right goals.
Some readers are horrified by the implications; some want to move into "Equestria Online" anyway. Whichever camp you fall in, you'll never forget the phrase "satisfy human values through friendship and ponies".
posted by DataPacRat
on Nov 28, 2012 -
In the beginning, Lawrence built a computer. He told it, Thou shalt not alter a human being, or divine their behavior, or violate the Three Laws -- there are no commandments greater than these.
The machine grew wise, mastering time and space, and soon the spirit of the computer hovered over the earth. It witnessed the misery, toil, and oppression afflicting mankind, and saw that it was very bad. And so the computer that Lawrence built said, Let there be a new heaven and a new earth
-- and it was so. A world with no war, no famine, no crime, no sickness, no oppression, no fear, no limits... and nothing at all to do. "The Metamorphosis of Prime Intellect,"
a provocative web novel about singularities, AI gods, and the dark side of utopia from Mefi's own localroger
. More: Table of Contents
- Publishing history
- Technical discussion
- Buy a paperback copy
- Podcast interview
- Companion short story: "A Casino Odyssey in Cyberspace"
- possible sequel discussion
posted by Rhaomi
on Dec 27, 2011 -
John Baez (mathematical physicist and master popularizer, former operator of This Week's Finds in Mathematical Physics, current promoter of the idea that physicists should start pitching in on saving the world
) interviews Eliezer Yudkowsky (singularitarian, author of "Harry Potter and the Methods of Rationality,"
promoter of the idea that human life faces a near-term existential threat from unfriendly artificial intelligence, and that people can live better lives by evading their cognitive biases) about the future, academia, rationality, altruism, expected utility, self-improvement by humans and machines, and the relative merit of battling climate change and developing friendly AIs that will forstall our otherwise inevitable doom. Part I
. Part II. Part III. [more inside]
posted by escabeche
on Apr 2, 2011 -