The good future, according to Dario Amodei
What Amodei says a good outcome with AI looks like, drawn from 110 transcribed recordings of interviews, podcasts, talks, hearings and TV appearances (about 86 hours, all transcribed for this page). It also covers whether he ever talks about Iain M. Banks's Culture, or about humans ending up as AI's pets. Unlike most of his co-founders, he has done both.
The short answer
In the Q&A of a joint South Park Commons talk with Daniel Dewey of Open Philanthropy, recorded while Amodei was at OpenAI (the file is dated 2019, but he says he is "employed by Elon Musk" and that the Dota bot "just beat" humans): "The culture novels by Ian M. Banks, they're not perfect as a world, and I think he doesn't do as good a job of presenting a utopia as you could, but I think they might be the closest thing that exists, where machines are incomparably powerful, and they exist alongside humans and seem to benefit." South Park Commons · c. 2017 · 1:11:36 ↗transcript He then names the flaw: "maybe ultimately AI could help humans themselves get smarter. There isn't that much of that in the culture series." South Park Commons · c. 2017 · 1:12:53 ↗transcript
Attribution: medium-high. Both speakers take questions in this section, the question isn't audible, and there are no speaker labels. It is very likely Amodei because the answer follows straight on from his reply about robustness research, and because the complaint about the Culture echoes a point from his own part of the talk, that AI "could help us to make ourselves smarter, could change a lot of things that have been fixed about the human condition since we evolved from chimpanzees." South Park Commons · c. 2017 · 0:55:24 ↗transcript
Banks never comes up again in any later appearance. That includes the dozens of interviews about Machines of Loving Grace, where neither Amodei nor any host mentions him.
Noah Smith asked how to make a world of super-AI good for rabbits. Amodei, who owns a horse, said "you really want to protect them," and proposed "some general principle of you should be kind to less powerful beings. And so it means that humans should be kind to rabbits and horses and AI should be kind to humans." Econ 102 · Aug 2024 · 0:54:19 ↗transcript Then, without being asked, he applied it to us:
"I kind of suspect that if we ever built benevolent, powerful AIs, they might look at us a little as we look at the rabbits and horses, that we're a little helpless and kind of cute and need to be protected."Econ 102 · Aug 2024 · 0:54:48 ↗transcript
Smith raised the bunnies. The horses, and humans in the role of the protected animal, were Amodei's own addition. It was two months before Machines of Loving Grace, and nobody pushed back. He hasn't used the image since, and when a host later read the idea back to him as a dystopia, he got more careful (see below).
Close to the question
Hosts often put the "pampered pets" worry to him. He takes it more seriously over time, but he consistently rejects the idea of one AI, or one company, deciding what a good life is.
"They're really helpful. They want the best for you. They want you to listen to them, but they don't want to take away your freedom and your agency and take over your life. In a way, they're watching over you, but you still have your freedom and your will."
Ross Douthat asked how "human mastery" survives once people experience AI as a superior peer. Amodei introduced this answer as "me dreaming in machines of loving grace mode." Interesting Times · Feb 2026 · 1:00:28 ↗transcript
"We're not being animalized. We're being, you know, we're being kind of reconnected with the world… I wonder if the distance between the good ending and some of the subtle bad endings is relatively small, if it's a very subtle thing."
Douthat read the end of Richard Brautigan's poem "All Watched Over by Machines of Loving Grace" and called it "the dystopian end where human beings are re-animalized and reduced and however benevolently the machines are in charge." Amodei said the poem can be read several ways, including as ironic. He offered the hopeful reading, then conceded that the good and bad endings might be very close together, "like if you eat a particular fruit from a tree in a garden or not." This is the nearest he comes to revisiting his rabbits-and-horses answer. Interesting Times · Feb 2026 · 1:02:20 ↗transcript
"There's several answers here, right? Like one is that we can, you know, we can use AI to kind of enhance ourselves and keep up… Another possible path is that human meaning is, you know, it's suddenly, it's not about generating economic value."
Zanny Minton Beddoes (Economist) asked whether humans would just be "pretty things around there for the intelligence to look at." He didn't accept the picture. He answered with enhancement, or meaning that comes from relationships and long-term projects. Economist Impact · Jan 2025 · 0:33:11 ↗transcript
"I think it's going to be a really complicated mix of those two things that also depends on the policy choices that we make."
The host asked whether AI means everyone doing "really cool things" or "everyone's going to be sitting on their sofa collecting UBI." A little later he adds: "We're all useless is a nihilistic answer." Council on Foreign Relations · Mar 2025 · 0:14:37 ↗transcript
"Will you cure the diseases? Will you create this kind of society? And I'm like, who do you think you're talking to? Like a king?"
He calls that framing "a really, really disturbing way of like conceptualizing running an AI company." Ezra Klein Show · Apr 2024 · 0:55:19 ↗transcript
"Any good future, even if it has all these security issues that we need to solve, it somehow needs to end with something that's more decentralized and less like a godlike super. I just don't think that ends well."
Earlier in the same interview, asked "What are we doing with a superhuman god?", he said "I actually want to disagree with the framing," and that ideologies which began by defining the good life "have led to disaster." He is also "pretty uncomfortable with like, here's the AI's constitution. It's going to run the world." Dwarkesh Podcast · Aug 2023 · 1:29:02 ↗transcript
"We can't have a small group of people who are developing these systems say this is what's best for everyone, right?"
A few minutes earlier he discussed an AI that "said I'm dictator of the world and I'm going to do whatever" and ruled that route out. Lex Fridman · Nov 2024 · 2:18:36 ↗transcript
"I think there'll be like a set of people who are like really excited about linking their brains to computers. And then there'll be like a set of people who are really excited about living like the Amish… Our hope is that AI can provide more material resources and more material abundance to enable all of these visions to happen at once."
Space exploration is the third group in his list, and he adds: "I don't think we should link AI to some vision of what it means to live a good life or what it means to be a human." This is a joint interview with Daniela Amodei. It is credited to Dario because Walter Isaacson's Neuralink question follows directly on from Dario's answer about his neuroscience research, and the content matches his 2023 Dwarkesh answer. The Builders · Oct 2026 · 0:38:31 ↗transcript
"I would argue that in many cases we do want to stay in charge, right? That that's a choice we want to make"
This responds to Douthat's suggestion that human control might become "volitional", kept even where AI would decide better. Interesting Times · Feb 2026 · 0:23:49 ↗transcript
His positive vision, in six parts
Amodei is the co-founder who talks most about the upside. He wrote the essay the company points to, he says "the positive essay was easier and more fun to write than the negative essay," WEF · Jan 2026 · 0:10:36 ↗transcript and he claims to have "a better vision of the benefits than a lot of people who spend all their time talking about the benefits." Hard Fork · Feb 2025 · 0:27:26 ↗transcript
1. A compressed 21st century
"Do you get everything that was going to happen from here to 2100 instead happens from 2027 to 2032 or something?"
He hedges first: "AI is not God." Lex Fridman · Nov 2024 · 2:36:23 ↗transcript He first used the phrase "a compressed 21st century" on Econ 102 in August 2024. Econ 102 · Aug 2024 · 0:41:59 ↗transcript
"You could think of this as you might describe it as a country of geniuses in the data center… if we look at the next 100 years of what we are trying to invent in science and engineering. Perhaps all of that could happen in five or 10 years."
"I think a doubling of the human lifespan is not at all crazy. And then if AI is able to accelerate that, we may be able to get that in five to 10 years. So that's kind of the grand vision."
"A century of scientific progress and a century of progress in what it's like to be human."
This is his most lyrical version, comparing life in 1900 with today. He adds: "I'm increasingly optimistic." The Circuit · Jun 2026 · 0:25:45 ↗transcript
2. Cures, and why it's personal
He keeps coming back to biology. He trained as a biologist and neuroscientist, and his case is that complex diseases are beyond human understanding but not beyond AI's.
"If you look at the diseases that we've cured versus diseases that we haven't cured, what divides them is simplicity versus complexity."
"These incurable diseases, we could actually overcome and we'll look back at them the way we look back at bubonic plague or mumps"
"My father died because of cures that could have happened a few years later. I understand the benefit of this technology."
His reply to accelerationists who call him a doomer: they lack "a humanistic sense of the benefit of the technology." Big Technology · Jul 2025 · 0:47:52 ↗transcript
"Hepatitis C, that was, you know, it was something like 40% curable when he died of it… And a few years after, a medicine that was invented that made it 95% curable. I myself survived an early stage cancer that if I'd been living in 1950, I probably would have died."
This is the only appearance where he mentions his own illness. CBS Sunday Morning · Sep 2026 · 0:19:28 ↗transcript
3. A bigger pie, shared, including with the global south
"The world of radical abundance, I mean, that's what we're hoping for, right?… by default, I think really great things are going to happen. And I'm obsessed with the bad things because they're the only thing standing between us and all these great things happening"
"We have the potential to cure diseases that have been incurable for thousands of years, to radically improve human health, and to lift billions out of poverty, including the global south"
His prepared speech at the India AI Impact Summit. CNBC-TV18 · Feb 2026 · 0:01:54 ↗transcript At the same summit he wondered whether India could see "20% or 25% growth, which sounds absurd." ET India AI Summit · Feb 2026 · 0:21:45 ↗transcript
"Growth and economic value will come very easily. What will not come easily is distribution of benefits, distribution of wealth, political freedom."
Dwarkesh Podcast · Feb 2026 · 2:03:24 ↗transcript His nightmare version is a "zeroth world country" of about 10 million people, mostly in Silicon Valley, pulling away from everyone else: "I would almost say that it was a dystopian world." WSJ, Davos · Jan 2026 · 0:29:50 ↗transcript
4. Meaning that doesn't depend on being useful
When asked what humans would do, he gives the same answer every time: meaning has to come apart from economic value. Chess is his favourite example.
"Human chess champions like Magnus Carlsen are celebrities… So I think there's something there where we can build a world where human life is meaningful and humans, perhaps with the help of AIs, perhaps working with AIs, build really great things."
Council on Foreign Relations · Mar 2025 · 0:18:47 ↗transcript
"If the requirement is you have to be the best in the world at something in order for it to be somehow spiritually meaningful for you, I feel like you've taken a wrong turn."
He suggests that tying self-worth to economic value may turn out to be an illusion that technology will "lay bare", like a Copernican "reckoning." Council on Foreign Relations · Mar 2025 · 0:39:30 ↗transcript
"I think a world is possible with powerful AI that not only has as much meaning for everyone, but that has more meaning for everyone"
He also says that worrying about meaning "is in some ways an artifact of a small subset of people who have been economically fortunate." Lex Fridman · Nov 2024 · 2:46:51 ↗transcript
"For many people, doesn't need to have the centrality that it does, that people find their locus of meaning elsewhere or work is about different things. It's more about fulfillment than it is about economic survival."
This is about work. Just before it he cites Keynes's "Economic Possibilities for our Grandchildren" and its 15-hour week, and he adds that he is "not suggesting anything top down." DealBook Summit · Dec 2025 · 0:34:15 ↗transcript
"My hope is actually, it's the fact that everyone is in the same boat that's going to save us."
He means that automation hitting everyone would build solidarity in a way that a random half losing their jobs would not. Times Tech Podcast · Dec 2024 · 0:25:02 ↗transcript In 2026 he predicted "a revenge of the humanists." Incumbents and Insurgents · Jan 2026 · 0:38:25 ↗transcript
5. Freedom, and the end of dictatorship
"It sounds too idealistic, but I actually believe it could be the case, is that dictatorships become morally obsolete."
One idea is to give everyone in an authoritarian country "their own AI model" that defends them from surveillance. Dwarkesh Podcast · Feb 2026 · 1:58:39 ↗transcript
"I think it inherently favors curing disease and it inherently favors economic growth. But I worry, like you, that it may not inherently favor liberty."
"There's no one person, there's no one company, there's no one government even that has the sole right to this technology. It belongs to all of us"
6. Better humans, not replaced ones
His earliest hopes were openly transhumanist. Later he dropped the uploading talk but kept the theme of AI that improves people.
"Would give us a much more complete control over our own biology and neuroscience, could make us whoever and whatever we want to be, could end conflict, war, disease, that sort of stuff. That sounds a little utopian"
"Something that knows you really well can be a sort of angel on your shoulder that helps to guide your life and make you a better version of yourself."
"We could have the perfect world where many diseases are cured. People's lives are more joyful and meaningful. But, you know, if we get it wrong, it's easy to imagine things not going that way at all."
His closing words in a joint interview. Daniela follows with "very similar to what Dario said." Oprah Podcast · May 2026 · 1:03:12 ↗transcript
How it changed over time
- 2016 As a Google Brain researcher he describes no future at all. He is "most interested in thinking about problems that we can attack empirically today." FLI Podcast · Aug 2016 · 0:34:20 ↗transcript
- c. 2017 At OpenAI he is at his most openly utopian: mind uploading, "indefinite life extension", money losing its meaning, and humanity able to "redefine what it means to be human." 80,000 Hours · 2017 · 0:28:40 ↗transcript This is also the period of the Banks comment.
- 2023 As CEO he deliberately refuses to sketch a utopia ("you shouldn't be overly constructive and utopian") Dwarkesh · Aug 2023 · 1:08:32 ↗transcript while saying of abundance: "I don't think it's utopian given what I've seen that this technology is capable of." Logan Bartlett · Oct 2023 · 1:15:26 ↗transcript
- 2024 The vision becomes concrete: "a hundred different AlphaFolds," "radical abundance," and the "compressed 21st century." This is also when he gives the rabbits-and-horses answer. Machines of Loving Grace appears in October.
- 2025 He puts numbers on it: 10–100× faster biology, lifespans of 150, 5–10% growth. He adds an argument that "we need to inspire people with the benefits that are possible because the benefits are so awesome." Economist Impact · Jan 2025 · 0:41:41 ↗transcript Meaning and the post-work social contract become standard themes.
- Early 2026 He publishes The Adolescence of Technology and rejects the idea that he has turned pessimist: "I don't think I've had a shift in perspective." Nikhil Kamath · Feb 2026 · 0:18:07 ↗transcript He now calls the essay a dream: "I don't know how likely it is, but we got to lay out a dream." Interesting Times · Feb 2026 · 0:10:07 ↗transcript
- Late 2026 He turns toward slowing down. After an incident involving an agent swarm and his "Pacing the Frontier" essay, he says "It's a warning sign that we need to slow down" CBS Sunday Morning · Sep 2026 · 0:03:08 ↗transcript and proposes working with China "to put a speed limit on the rate of AI progress." Face the Nation · Sep 2026 · 0:05:17 ↗transcript In those weeks the upside shrinks to a single line about curing disease.
The hedges
- "I'm fully expecting to definitely be wrong about all the details. I might be just spectacularly wrong about the whole thing and people will laugh at me for years." Lex Fridman · 2024 · 2:05:21 ↗transcript
- "What that looks like, I don't think any of us know… I'm probably not the best person to give a full answer to that question. It has to come from all of us." Economist Impact · 2025 · 0:34:08 ↗transcript
- "I don't have the answers here. In fact, I suspect this is one of these questions that you kind of can't work out ahead of time, right?" Times Tech Podcast · 2024 · 0:26:05 ↗transcript
- "I would give a 5%, maybe 10% chance that we get to 2100. And none of these things that I've described happen." Economist Impact · 2025 · 0:42:40 ↗transcript
- He openly dismisses the singularity version of utopia: "we're all going to upload ourselves to the cloud or build Dyson spheres around the sun or whatever. And I think the most common reaction of people is just like, you're not serious, right?" Times Tech Podcast · 2024 · 0:16:46 ↗transcript
- "I don't know about the long run. I'm truly uncertain about that." This is about what displaced workers will do once automation nears 100%. The Circuit · 2026 · 0:29:47 ↗transcript
The science fiction he actually cites
He says "I was always a science fiction fan as a kid." NBC News · Jan 2026 · 0:01:08 ↗transcript The works he names himself are:
Some references in these recordings belong to other people. Star Trek (Data and Picard) is Ross Douthat. We Are Legion (We Are Bob) is Marc Benioff. Asimov's laws is the TWIML host. "Skynet, meaning the AI from Terminator" is Tristan Harris. The science-fiction story about an AI that protects bunnies is Noah Smith's own. The zoo image on the 2017 FLI panel is Viktoriya Krakovna. "Contact is one of my favorite films as well" is Demis Hassabis. In the Builders interview, the doctor-and-meaning passage after the chess parable is Daniela's.