bayesing
Original: bayesing on Saturday Morning Breakfast Cereal
Transcript
Panel 1:
Woman with gray hair (off-panel, left): I think we should all try to be Bayesians.
Woman with dark curly hair and round glasses: What?! Bayesianism can't possibly be true, and the proof is easy.
Panel 2:
Woman with glasses: Suppose your prior probability that 0 equals 1 is precisely 0.
Panel 3:
Woman with glasses: You take a field observation and find that to your surprise 0 equals 1.
Panel 4:
Woman with glasses: The probability of that observation, given your prior assumption, is, let's say, 0.5.
Panel 5:
Woman with glasses: Bayesians should update, but they can't because the observation that 0 equals 1 would change the mathematics of Bayes' Theorem directly, making the probability both 0 and 1!
Panel 6:
Woman with gray hair: Your argument is that Bayes' Theorem can't account for incredibly stupid shit?
Woman with glasses: And thus cannot be a complete description of our universe!
Votey:
Three people sit together, one (with frizzy hair) grinning at a laptop/screen. Thought bubble: People love relatable humor
Woman with gray hair (off-panel, left): I think we should all try to be Bayesians.
Woman with dark curly hair and round glasses: What?! Bayesianism can't possibly be true, and the proof is easy.
Panel 2:
Woman with glasses: Suppose your prior probability that 0 equals 1 is precisely 0.
Panel 3:
Woman with glasses: You take a field observation and find that to your surprise 0 equals 1.
Panel 4:
Woman with glasses: The probability of that observation, given your prior assumption, is, let's say, 0.5.
Panel 5:
Woman with glasses: Bayesians should update, but they can't because the observation that 0 equals 1 would change the mathematics of Bayes' Theorem directly, making the probability both 0 and 1!
Panel 6:
Woman with gray hair: Your argument is that Bayes' Theorem can't account for incredibly stupid shit?
Woman with glasses: And thus cannot be a complete description of our universe!
Votey:
Three people sit together, one (with frizzy hair) grinning at a laptop/screen. Thought bubble: People love relatable humor
Alt text
Main comic, six panels in a green setting. A gray-haired woman tells a dark-haired woman with round glasses, "I think we should all try to be Bayesians." The glasses woman replies, "What?! Bayesianism can't possibly be true, and the proof is easy." She then lays out an absurd proof over several panels: "Suppose your prior probability that 0 equals 1 is precisely 0. You take a field observation and find that to your surprise 0 equals 1. The probability of that observation, given your prior assumption, is, let's say, 0.5. Bayesians should update, but they can't because the observation that 0 equals 1 would change the mathematics of Bayes' Theorem directly, making the probability both 0 and 1!" The gray-haired woman deadpans, "Your argument is that Bayes' Theorem can't account for incredibly stupid shit?" The glasses woman raises a finger triumphantly: "And thus cannot be a complete description of our universe!" Votey: a rough sketch of three people sitting together, one grinning at a laptop with a thought bubble reading "People love relatable humor."
Transcribed by Claude Opus 4.8.