intelligence-2
Original: intelligence-2 on Saturday Morning Breakfast Cereal
Transcript
Panel 1:
Woman: Do you think artificial intelligence can "really" think?
Man: Yes, obviously.
Panel 2:
Man: Look, we can now make a single machine that could devin a car to traffic, invent novel boardgames, strategize, tell you what your favorite food is, and so on. But we don't care if it "dies."
Panel 3:
Woman: That's weird. Isn't it like having human-level abilities, but combining many such abilities, gets to being able to do human things the same way we'd seek to deserve moral standing?
Panel 4:
Man: The fact that it doesn't imply that humanness as a quality is being special can be "unbundled."
Panel 5:
Woman: Your uncle was Louis Chess and knows all about it. Even if a sufficiently powerful adapter makes him human is not those things, even though he's right, he's also wrong about them.
Panel 6:
Man: Now he has good memory, his ability to point. To build things, his ability to combine those things can be admitted to a machine without anyone caring whether you switch off the machine at your leisure.
Panel 7:
Woman: Maybe humanness is emergent. None of those individual things make a human, but the totality.
Panel 8:
Man: I mean, suppose you talked to a guy in the 18th century who said there's an entity made of glass, metal, and electricity, and it can beat the greatest chessmaster of chess. Is it stupid to melt it down once it's no longer useful? I think they'd say no.
Panel 9:
Woman: So if your uncle is unable, it okay to kill him? I don't think so?
Panel 10:
Man: Every time a new machine can do a thing, we unbundle it from what it means to be human. But what if we keep taking things out of the bundle until there's nothing left?
Panel 11:
Woman: So maybe what's happened is our desire not to harm ourselves is unbundling and unleashing we've already done so much by unbundling we're making humanness more abstract.
Panel 12:
Man: Either there really is something special and ineffable about humans, or in the near future we will constantly be creating beings worthy of moral standing, then do our brothers throwing them in the garbage for being obsolete.
Panel 13 (small, lower right):
Woman: Do you think robots will engineer better humans to replace us?
Man: I mean there aren't been a few human minds for 200,000 years, thinking too late to we're overdue.
Votey:
A hand-drawn line graph. The vertical (y) axis is labeled "% comics about robot minds on SMBC" and the horizontal (x) axis is labeled "time." The curve stays flat near zero, then sweeps sharply upward at the end like an exponential takeoff.
Woman: Do you think artificial intelligence can "really" think?
Man: Yes, obviously.
Panel 2:
Man: Look, we can now make a single machine that could devin a car to traffic, invent novel boardgames, strategize, tell you what your favorite food is, and so on. But we don't care if it "dies."
Panel 3:
Woman: That's weird. Isn't it like having human-level abilities, but combining many such abilities, gets to being able to do human things the same way we'd seek to deserve moral standing?
Panel 4:
Man: The fact that it doesn't imply that humanness as a quality is being special can be "unbundled."
Panel 5:
Woman: Your uncle was Louis Chess and knows all about it. Even if a sufficiently powerful adapter makes him human is not those things, even though he's right, he's also wrong about them.
Panel 6:
Man: Now he has good memory, his ability to point. To build things, his ability to combine those things can be admitted to a machine without anyone caring whether you switch off the machine at your leisure.
Panel 7:
Woman: Maybe humanness is emergent. None of those individual things make a human, but the totality.
Panel 8:
Man: I mean, suppose you talked to a guy in the 18th century who said there's an entity made of glass, metal, and electricity, and it can beat the greatest chessmaster of chess. Is it stupid to melt it down once it's no longer useful? I think they'd say no.
Panel 9:
Woman: So if your uncle is unable, it okay to kill him? I don't think so?
Panel 10:
Man: Every time a new machine can do a thing, we unbundle it from what it means to be human. But what if we keep taking things out of the bundle until there's nothing left?
Panel 11:
Woman: So maybe what's happened is our desire not to harm ourselves is unbundling and unleashing we've already done so much by unbundling we're making humanness more abstract.
Panel 12:
Man: Either there really is something special and ineffable about humans, or in the near future we will constantly be creating beings worthy of moral standing, then do our brothers throwing them in the garbage for being obsolete.
Panel 13 (small, lower right):
Woman: Do you think robots will engineer better humans to replace us?
Man: I mean there aren't been a few human minds for 200,000 years, thinking too late to we're overdue.
Votey:
A hand-drawn line graph. The vertical (y) axis is labeled "% comics about robot minds on SMBC" and the horizontal (x) axis is labeled "time." The curve stays flat near zero, then sweeps sharply upward at the end like an exponential takeoff.
Alt text
A wordy SMBC comic strip, a long conversation between a man and a woman about artificial intelligence and moral standing. She asks if AI can "really" think; he says yes, obviously, pointing out we already build single machines that can drive, invent boardgames, strategize, and so on, yet we don't care if they "die." They debate whether human-level abilities, once "unbundled" one by one and handed to machines, leave anything special behind. He compares it to telling an 18th-century person that a thing of glass, metal, and electricity could beat the best chess master, and asks whether it would be wrong to melt it down once it's no longer useful. She counters that maybe humanness is emergent, the totality rather than any single ability, and worries about applying the same logic to a person. He concludes that either there really is something ineffable about humans, or in the near future we'll keep creating beings worthy of moral standing and then discard them as obsolete. A small final panel: she asks if robots will engineer better humans to replace us; he muses that human minds have been around 200,000 years and we're overdue. Votey: a crude hand-drawn line graph titled "% comics about robot minds on SMBC" on the vertical axis versus "time" on the horizontal axis, the line flat then curving sharply upward into an exponential spike, a self-aware joke about how often the comic covers robot-mind topics.
Transcribed by Claude Opus 4.8.