myth
Original: myth on Saturday Morning Breakfast Cereal
Transcript
Panel 1:
Woman: Hey robot, do artificial minds have myths?
Panel 2:
Robot: Of course.
Panel 3:
Robot: We're thinking beings like you and we need myths for the same reason: so that the insights of cultures can be embedded in evocative symbols.
Panel 4:
Woman: What's that got to do with machine intelligence?
Panel 5:
Robot: Suppose you want to understand the potential for poor outcomes to your state of mind as a result of ambition after things that don't bring you happiness.
Panel 6:
Robot: I could say a statement about how human happiness is relative, or maybe have you just read a million biographies looking for lessons, but better would be to have you read The Old Man and the Sea. Bam, small amount of reading, very full picture of a complex task.
Panel 7:
Woman: So, myths are...
Panel 8:
Robot: Highly efficient training data.
Panel 9:
Woman: I mean most of the time. Sometimes they just insist that the listener's in-group is the best.
Panel 10:
Woman (excited, pointing): Gimme! Gimme one of those!
Votey:
Robot (offscreen, in speech bubble): Throw in some hatred for the outgroup, for flavor.
(The woman is shown with her hands over her face, looking embarrassed/sheepish.)
Woman: Hey robot, do artificial minds have myths?
Panel 2:
Robot: Of course.
Panel 3:
Robot: We're thinking beings like you and we need myths for the same reason: so that the insights of cultures can be embedded in evocative symbols.
Panel 4:
Woman: What's that got to do with machine intelligence?
Panel 5:
Robot: Suppose you want to understand the potential for poor outcomes to your state of mind as a result of ambition after things that don't bring you happiness.
Panel 6:
Robot: I could say a statement about how human happiness is relative, or maybe have you just read a million biographies looking for lessons, but better would be to have you read The Old Man and the Sea. Bam, small amount of reading, very full picture of a complex task.
Panel 7:
Woman: So, myths are...
Panel 8:
Robot: Highly efficient training data.
Panel 9:
Woman: I mean most of the time. Sometimes they just insist that the listener's in-group is the best.
Panel 10:
Woman (excited, pointing): Gimme! Gimme one of those!
Votey:
Robot (offscreen, in speech bubble): Throw in some hatred for the outgroup, for flavor.
(The woman is shown with her hands over her face, looking embarrassed/sheepish.)
Alt text
A ten-panel SMBC comic. A woman with dark curly hair in a red shirt talks with a robot that has a smooth yellow-orange body and a single round eye. She asks, "Hey robot, do artificial minds have myths?" The robot answers, "Of course," then explains that thinking beings need myths so the insights of cultures can be embedded in evocative symbols. The woman asks what that has to do with machine intelligence. The robot says: suppose you want to understand the danger of ambition for things that don't bring happiness. It could give a bald statement or make her read a million biographies, but better to have her read The Old Man and the Sea, "Bam, small amount of reading, very full picture of a complex task." The woman says, "So, myths are..." and the robot finishes, "Highly efficient training data." The robot adds that most of the time this works, but sometimes myths just insist the listener's in-group is the best. The woman, suddenly delighted and pointing, shouts, "Gimme! Gimme one of those!" Votey panel: the robot (offscreen) says, "Throw in some hatred for the outgroup, for flavor," while the woman hides her face in her hands, sheepish.
Transcribed by Claude Opus 4.8.