17 | Annalee Newitz on Science, Fiction, Economics, and Neurosis
What this covers
Annalee Newitz and Sean Carroll explore what science fiction actually does and why it matters. Newitz, who founded io9 and has recently turned to novel writing with Autonomous, argues that science fiction's real work is neither prophecy nor speculation but the construction of economically and socially coherent imagined worlds. The conversation ranges across her new novel—which asks what happens when the right to work becomes a commodity rather than a birthright, and when synthetic drugs and sentient robots upend pharmaceutical economics—and touches on how the scientific method appears in fiction, what real intelligence might look like, and why the humanities matter to our understanding of technology.
Throughout, both speakers return to a central claim: that human behavior, like machine behavior, emerges largely from coping with constraints, neuroses, and inherited cultural programming rather than from rational optimization. Where Asimov's psychohistory breaks down is precisely here—humans interact nonlinearly, tiny variations amplify, and historical trauma does not dissipate but echoes forward in new forms. This shapes how economies work, how AI will actually integrate into society, and what makes world-building believable. The conversation touches on why machine learning differs from strong AI, why capitalism persists even in imagined futures, and how intelligence itself might be better understood as an ecosystem of different kinds rather than a single achievable goal. Newitz emphasizes that incorporating economics, labor, and historical struggle into science fiction narratives is not a detour from the story—it is what makes the story matter to readers who live inside those same pressures.
Newitz argues that science fiction's job is not to predict the future but to build economically and socially realistic imagined worlds, and that doing so well requires treating economics, history, and human neurosis as central rather than incidental to stories about science and technology.
- Science fiction needs more economics because economics shapes how science actually gets done and how people live
- Human and machine behavior alike comes down largely to coping with neuroses and cultural programming
- Realistic world-building requires understanding recurring human patterns, not predicting specific futures
Economic constraints, property regimes, and labor systems recur in new forms across centuries and will persist even in speculative futures.
- In Autonomous, the UN has been replaced by an International Property Coalition where human rights are superseded by and conceived through property rights; robots are human-equivalent yet born indentured to pay off their manufacturing cost, and through legal shenanigans humans too become subject to 'human rights indenture laws' allowing them to sell themselves into indenture—a reinvention of slavery dressed in euphemism.
“human rights have been superseded by property rights, and that's how people think of human rights, is in the context of property... robots are human-equivalent in this world and they are born indentured to pay off the cost of their manufacture... humans also are now subject to what they call the 'human rights indenture laws'.”
- Pharmaceutical companies are a huge, poorly understood business whose money-making overlaps heavily with the equally poorly understood patent system, making it an easy way to exploit a captive audience—since no one refuses a drug that cures or prolongs their life—which made it a natural target and the source of real anxiety and pain for many people.
“pharma companies are huge, it's a huge business... the way that they make money overlaps a lot with the patent system, which is also poorly understood. And so it's a great way to just rip people off, and you always have a captive audience. Nobody's gonna say no to a drug that's gonna cure them or prolong their life.”
- Even in the present world there are jobs that amount to indenture—graduate school being a prime example—where, especially in the humanities, you effectively pay into the university system in order to have the right to work there, which made it plausible that future characters would choose self-indenture.
“even in our world now, there's plenty of jobs that people do that are basically indenture. Well, like being a grad student, for example... It's the university system, is you pay in in order to have the right to work there, at least in my experience.”
- In 150 years we will not have gotten rid of capitalism, and science will not magically pay for itself; future people, even with hard AI, will still be struggling to get grants and still wrestling with social inequality, because economic constraints and historical struggles recur in new forms rather than vanishing.
“I don't think in 150 years, we're gonna have gotten rid of capitalism and we're not gonna be... Science is not magically gonna pay for itself. And so I really wanted to think about how people in the future would still be... still trying to get grants and they're still struggling with social inequality.”
- The rights we take for granted—such as the right to work anywhere or live anywhere—have nothing to do with capitalism and are merely artifacts of how a society was set up; they are amazing rights to possess at birth that many civilizations have never had, and they can be taken away and replaced by paid 'property' versions if the economic system changes.
“we think that we should just be allowed to live wherever we want. But actually that's a really amazing right to just have that at birth. A lot of civilizations didn't have that and don't have that.”
Intelligence is a spectrum of diverse cognitive types, not a singular human standard, and artificial minds could replicate human neuroses.
- Rather than a robot uprising, AI and machine learning will merge into human society in ways often hard to tell apart from it—via implants and partial AI—producing a spectrum from totally biological human intelligence to totally machine intelligence with everything in between.
“I don't think we're gonna have a robot uprising. I think it's much more like robots and AI and machine learning will kind of merge into human society in ways that are sometimes hard to tell apart from human society... we're gonna end up with a spectrum where there's totally biological human intelligence... all the way on the other side is like totally machine intelligence and everything in the middle.”
- As machine learning advances, we are realizing we do not actually know what intelligence is in humans or anything—we are at square one—and trying to recreate something we cannot even quantify in ourselves is extremely hard, which undermines confident attempts to build human-equivalent AI.
“we don't know what the hell intelligence is at all in humans, in anything. We are totally square one with that... trying to recreate something that we can't even quantify in ourselves is really tough.”
- An artificially intelligent robot like Paladin could go into her own mind, see all the programs she is running, identify why she feels a certain way, and selectively modify those programs—something humans cannot do with their own cultural conditioning—which dramatizes the idea that humans, too, are physical organisms running 'programs' we cannot simply excise.
“for a robot, she actually can go in eventually and see all the programs she's running and be like, 'Oh that's why I was feeling that way. Oh, I could modify that'. So robots would be great in therapy.”
- Because humans live in 'meat sacks' with built-in urges to eat and reproduce, culture has swarmed over these simple biological activities and, via what Foucault calls a 'perverse implantation,' trained people to regard their own desires as terrible or wrong, fetishizing eating and sex with infinite ritual complexity—and one could similarly program an AI to have the same fucked-up, neurotic desires.
“using culture, we can try to train people that these urges and needs are bad... That's why people have so many weird rituals around sex and food because these are these basic things that we want and culture has just swarmed all over them and messed them up... we fetishize them to the point of absurdity.”
- Intelligence is best understood as an ecosystem of different kinds working on different problems—some appearing unintelligent yet brilliant at particular tasks—and recognizing that neurotypicality is not actually typical helps us define intelligence as a spectrum of consciousness rather than one recognized type with everything else labeled deviant.
“intelligence to the extent that it can be talked about is really an ecosystem. There's different kinds of intelligence... once we have an idea that there's a spectrum of what it means to have consciousness and intelligence, and it's not just like, here's the one kind of consciousness that we recognize as intelligent and everything other than that is like autistic or a schizophrenic”
Science fiction and serious fiction must portray economics, labor, and scientific process accurately to engage readers and reflect real human struggle.
- The transition from academic writing to popular writing was actually much harder than the transition from science nonfiction to science fiction, because academic writing relaxes the demand for mass appeal and lets you be quirky for a tiny expert audience, whereas popular writing requires the learned craft of engaging readers and telling a relatable story.
“making the transition from academic writing to popular writing, I have to say, was actually much more difficult in many ways than transitioning from science nonfiction to science fiction... Part of your job is also just to engage the reader which... It's a whole craft that you have to learn.”
- Studying computer security and hacking is a great way to learn about computer networking because it concerns how subversive activities sneak around supposedly safe protocols; it is almost like studying 'computer network neurosis'—how unwanted thoughts and applications appear in the otherwise pure mind of a networked computer.
“it's almost like studying computer security or writing about it, it's almost like looking at computer network neurosis because it's all about how to... Like unwanted thoughts and applications appear in this good pure mind of the computer or mind of your network of computers.”
- Science fiction (and fantasy) should incorporate the economic dimensions of its scenarios—funding, grants, property rights, labor—because economics consumes the actual lives of scientists and is a big part of how science is done, yet these wonky-sounding questions are precisely what make stories exciting since they show people struggling to get what they want.
“we kind of see characters grappling with the economic dimensions of what's going on, whether that's space travel and colonizing other worlds, or whether that's getting a grant to study something, which we almost never see in science fiction. And it's like it consumes the lives of scientists. It's actually a big part of how you do science.”
- The most important thing in representing science in fiction is portraying the scientific process accurately—people testing and retesting things repeatedly—rather than depicting science as gazing rapturously into glowing holes, even when the underlying technology (like photonic molecules) is invented, because science is mostly dreary repeated testing.
“The thing that's most important to me when representing science is trying to represent the scientific process accurately... science is actually a lot of dreary testing and not a lot of gazing rapturously into glowing holes and saying, 'It's full of stars'.”
- For Newitz, novels come together by starting with characters and the world they inhabit, focusing on character to write the book, and only returning at the end to smooth out and make the world's rules consistent—because info dumps bore readers and inconsistencies make every nerd wince, so most world-building is painted faintly in the background.
“it starts with characters, and I start to think about the world that they're in... then I have to come back again at the end and say, 'Okay, so what are the rules of the world? I need to smooth this out and make sure everything is consistent.'”
Livable futures should limit drama to interpersonal conflict rather than structural violence and killing.
- Across both her fiction and nonfiction, Newitz's recurring aim is to urge people to survive in a way that helps many others survive with them—showing how human cultures live on through catastrophe and how new social structures can be built that are less inclined to encourage killing.
“in all of my work, I have a strong wish to urge people to survive, and try to survive in a way that helps lots of other people survive with them.”
- The humanities and social sciences bring to science a sense of history and an ability to analyze social data—understanding how humans behave over time and under different conditions—and although human behavior produces recurring patterns, those patterns are not perfectly reproducible the way an ideal scientific experiment would be.
“what humanities and social sciences bring to science are a sense of history, and also an ability to analyze social data... we actually need to understand how humans act over time and what humans have done at various points... And there are patterns. They're not... It's not perfectly reproducible as an ideal scientific experiment would be.”
- Autonomous is neither a dystopia nor a utopia but a 'topia'—just how things are now, where some things are a little worse and some a little better for certain groups; even with an awesome Gini coefficient and no extreme rich-poor divide, humans will still argue, feel jealousy, and grieve, so the goal is not eliminating drama but limiting it to interpersonal conflict rather than wars and killing.
“I don't think we're ever gonna reach a point as a civilization or a species where we're happy all the time, even if we have an awesome genie coefficient and we don't have an extreme division between rich and poor... there's always gonna be drama and sadness. It's just that, hopefully, we can keep it... limited.”
Historical trauma and complex systemic failures propagate nonlinearly across time and cannot be undone by single progressive events.
- Major historical disturbances like the slavery of Africans in the US do not simply dissipate; their ripple effects propagate across centuries and, contrary to the hope that they would be cancelled out by events like the Obama presidency, the effects sometimes grow bigger over time or return in new, equally pernicious forms—mirroring the nonlinear amplification of fluctuations in complex systems.
“the kind of ripples that have propagated out from that historical trauma just don't... you would wish that they would just kind of bump into other things and that those ripples would bump into the Obama presidency and be like, 'A-ha everything's fixed', but it isn't fixed. And in fact, sometimes the effects get bigger over time, or they return in new ways that are just as pernicious.”
- Cities like Angkor are abandoned through a combination of factors: politics are a huge part of what causes a city to fail, but politics alone won't do it—it must be combined with an environmental problem such as climate change or mismanaging the water supply, plus recurring issues like historical trauma and infrastructure upkeep.
“Politics are a huge part of what causes a city to fail, but politics alone won't do it. It's really got to be combined with some kind of environmental problem, whether the problem is actually climate change or mismanaging the water supply or whatever.”