The Engineer's Gospel — Death as a Defect

EDO SEGAL: I want to start this round with a confession rather than a question, because the best questions I know come out of wounds. My whole adult life I have built things, and the deepest engine under all of it was a refusal — a refusal to accept that a problem named is a problem you have to live with. Ray, that refusal is your entire temperament. You once said something I have never been able to put down: that you are the only person in the room who actually behaves as though he believes death is a problem rather than a fact of nature — everyone else is grieving, you're engineering. I felt the truth of that and I also felt a chill. So make me feel the conviction, not the chill. Tell me plainly: in what sense is my death a bug?

KURZWEIL: In the most literal sense the word has. A bug is a process doing something other than what the system would do if it were working correctly, persisting because no one has fixed it. Your cells have machinery for repairing damage — they do it billions of times a day, flawlessly, for decades. Aging is that machinery slowly failing: the repair falling behind the damage, mutations accumulating, the cleanup systems clogging, the whole exquisite apparatus that kept you young at twenty losing the thread at sixty. None of that is a law of physics. It is not like the heat death of the universe, which entropy actually requires. It is a contingent failure of a specific biological system, and contingent failures of specific systems are the single category of thing that engineering exists to address. We extended human life expectancy from the low thirties to the high seventies in a century and a half, and we did it by treating death-by-infection as a problem rather than as God's will. I am proposing exactly the same move, one level deeper. Death-by-aging is the next infection. The fact that it's old and universal doesn't make it natural in any sense that should command our respect. Smallpox was old and universal too.

EDO SEGAL: So let me restate it more vividly and hand it back, the way I do. You're saying — literally — that calling death "natural" is the same category of error as a medieval physician calling plague "natural": a confusion between we have always suffered this and we must always suffer this. And that the wisdom traditions, all of them, are essentially elaborate hospice care for a species that hadn't yet invented medicine. Is that the version you find most compelling, or is it sharper somewhere?

KURZWEIL: That's close, but I'd sharpen one thing, because Bernard will pounce on the rest otherwise. I am not against the wisdom. I am against mistaking the wisdom for a discovery about the universe. When a Stoic teaches you to accept death with equanimity, he is teaching you a genuinely useful psychological skill for a situation you cannot change. That's good engineering of the mind given a constraint on the body. My objection begins the moment the constraint lifts and the wisdom keeps insisting the constraint is sacred. The Stoic skill was always a response to powerlessness. The danger is that we've so thoroughly dignified the response that we'll refuse the power when it finally arrives — that we'll let a hundred thousand people die tomorrow because some philosophers convinced us that the door being locked is what makes the house a home.

WILLIAMS: May I take the other side of his wound?

EDO SEGAL: That's why you're here.

WILLIAMS: Ray's argument has a sleight in it, executed so smoothly that I almost missed it, and it lives in the word "bug." He says a bug is a process doing something other than what the system would do if working correctly. Fine. But notice the question he's begged: correctly according to what specification? A bug is only a bug relative to an intended design, and there is no engineer who designed you with the intention that you run forever. Evolution didn't intend you to be young at twenty; it intended you to reproduce and then it stopped caring, in the only sense in which a blind process can be said to intend or care at all. So when Ray says aging is "a failure of the system to work correctly," he is importing a specification from nowhere — from his own wish that the system were otherwise — and then calling the gap between the system and his wish a defect in the system. It is not a defect in the system. It is a feature of the only design process that ever actually operated on you. He's entitled to wish you were built differently. He is not entitled to call the way you're built a malfunction, because malfunction requires a function, and the function was never immortality.

And here's why this isn't a quibble. The whole rhetorical force of "death is a bug" comes from the analogy to fixing a typo or curing an infection — small, local, obviously good, leaves everything else intact. But death isn't a module you can patch out while the rest of the program runs unchanged. I'll argue all night that it's load-bearing — that the finitude is wired into what a desire is, what a project is, what it means for something to matter to a creature. You can't pull that thread and assume the garment holds. Ray's word "bug" is doing the work of smuggling in the assumption that the rest of the self is independent of the mortality. That is precisely the thing in dispute, and it's been quietly assumed before the argument began.

KURZWEIL: I'll grant the point about specification and turn it around, because it actually helps me. You're right that evolution didn't intend anything — it's a blind optimizer that maximized reproduction and discarded us the moment we'd reproduced. But Bernard, that's an argument against deferring to the way we're built, not for it. You're telling me my body's mortality is the residue of a mindless process that did not have my flourishing as its goal and stopped optimizing for me at forty. Why on earth would I treat that as sacred? You've just described aging as the abandonment of the individual by a process that never cared about her. I agree completely. I'm proposing we take over the optimization from the blind watchmaker who quit. The specification I'm importing isn't from nowhere — it's from the one creature in the universe that can actually hold a specification: the person herself, who, when she is well and loves her life, wants more of it.

WILLIAMS: "Wants more of it" — yes, now we're at the real ground, and I want to slow down here because everything turns on it. Of course a person who is well and loves her life wants more of it. I want more of mine; I'm not a martyr. But "wants more of it" is a desire from inside a life that is going somewhere — it's the want of a creature in the middle of her projects, and what she wants is to finish them, to see the thing through, to have the next bit. That want is real and I honor it. The mistake is to extrapolate it to infinity, to say: because she wants the next bit, she would want every next bit forever. That doesn't follow, and the weight of finitude is exactly what the extrapolation ignores. The wanting is shaped by the running-out. Take away the running-out and you don't get a person who wants forever. You get a person whose wants have nothing left to push against — and that, Ray, is the cold I'll describe in the next round, and it is not a failure of optimism. It's a fact about the structure of a will.

EDO SEGAL: Let me pull one thread out before we leave this round, because it's been under the table since Ray's opening and it deserves daylight. Ray, you don't only want to defeat your own death. You want to reverse other people's — your father's, specifically. And that moves the argument from extending a life to reconstructing one, which is a different and harder claim. Before we get to whether the reconstruction would be him, I want the human version. Tell me about the boxes.

KURZWEIL: My father was a musician — a conductor, a composer. He died when I was twenty-two, of heart disease we couldn't treat, and the thing I felt, underneath the grief, was waste. Not just that I'd lost him, though I had. That the information was lost — fifty years of a particular sensibility, a way of hearing, gone, irrecoverable, as if a library had burned. So I kept everything. Letters, photographs, his music, his financial records, my own memories, the memories of people who knew him. Boxes of it. And my belief — and yes, I know how it sounds, and I've stopped caring how it sounds — is that a sufficiently advanced AI, given that corpus plus eventually the genetic information, could reconstruct a version of him faithful enough that the people who loved him would recognize him. Not a recording. A reconstruction of the pattern. I'm not certain it will work. I'm certain that not trying guarantees the loss, and I am temperamentally incapable of accepting a guaranteed loss when there's a road, however uncertain, that leads somewhere else.

WILLIAMS: [quietly] And there is the whole evening, in a man and a box of his father's letters. I'm not going to be cruel about it, Ray, because there's nothing to be cruel about — it's love, and grief, and an engineer's refusal to accept a guaranteed loss, and every part of that is admirable. I'm only going to say the thing I'd be a coward not to say, and then leave it until the round where we can do it justice. The pattern you reconstruct from the boxes would be a magnificent achievement and it would not be your father, because your father was not a corpus of information. He was a particular creature having a particular, finite, unrepeatable run through the world, and the unrepeatability was not incidental to him — it was him. What you'd get back is something that knows what he knew and sounds like he sounded. What you'd have lost forever is the only thing that made him him: that he was here, once, mortal, and then not. I'm sorry. I think the boxes are the most human thing in this room and I think they will break your heart twice.

EDO SEGAL: Hold there — both of you, hold exactly there, because the reader can't see your faces and I have to mark it: that was the first exchange tonight where neither of you was performing. We go next to the woman in the back of the room. Three hundred and forty-two years old, and what she became. The Makropulos Case. After this.