The good future, according to Jared Kaplan
What Kaplan says a good outcome with AI looks like, drawn from 24 transcribed talks, lectures, panels and interviews. It also covers whether he ever talks about Iain M. Banks's Culture, or about humans ending up as AI's pets. The short version: he says very little about the good future, and what he does say is mostly about science.
The short answer
Banks, the Culture and benevolent "Minds" appear in none of the 24 appearances. We checked keywords and likely mistranscriptions, and read every transcript in full. When a host asked him to fill in the optimistic picture of Dario Amodei's Machines of Loving Grace, he talked about drug discovery and then said "In terms of how exactly it will roll out, I really don't know. It's really, really hard to predict the future." Y Combinator AI Startup School · Jun 2025 · 0:23:41 ↗transcript
He never talks about pets, zoos or house cats. His nearest point is a plain statement about losing power: "As humans, we're sort of used to being the smartest species on the planet. And so if we aren't anymore, then that really kind of changes the balance of power." Strings 2024, CERN · Jun 2024 · 0:03:28 ↗transcript When Azeem Azhar suggested AI could make us look "like hamsters", Kaplan left the image alone and answered about interpretability (see below).
Close to the question, but not quite
He never uses pet imagery, but he does talk about humans being outclassed, losing control and becoming dependent. He treats these as risks to manage, not as a future to accept.
"Yeah, it's a good question. So there are a couple of things I would say. So one is that if you want to understand everything that the AI is doing, I agree that's going to be very difficult."
Azhar asked what interpretability can mean once machines work "in spaces that make us look not so much like silverback gorillas, but like hamsters." The hamster image is Azhar's. Kaplan didn't take it up. He conceded the point and moved on to technical answers: interpretability as a way of "auditing AI", and using AI to supervise AI so that oversight improves "with the intelligence of AI." (A short stretch of his answer is missing from the transcript.) Azeem Azhar's Exponential View · Apr 2025 · 0:33:54 ↗transcript
"That sounds like me, yeah."
Earlier in the same interview, Azhar read out a Kaplan quote: "once you make a thing that's broadly much smarter than you, and given that it'd be easy to run millions of copies of that thing once you have one, you're going to lose and be disempowered if there's a conflict… being 90% sure it'll work out is very far from okay." Kaplan confirmed it was his. Azeem Azhar's Exponential View · Apr 2025 · 0:30:00 ↗transcript
"Should we be sort of having these super intelligent AI aliens kind of invading the earth or should we decide not to?"
He presents whether to build superintelligence at all as a real choice for society. Azeem Azhar's Exponential View · Apr 2025 · 0:47:10 ↗transcript
"In the modern world, I don't understand how most things work. Like, do I understand how my car works? Do I understand how my iPad works? So when you get more and more systems that no one really understands, things can kind of go off the rails."
This is about an economy of AIs dealing with other AIs ("your AI doctor interacts with your AI pharmacist"). It's the closest he gets to the idea of humans gradually losing their grip. Azeem Azhar's Exponential View · Apr 2025 · 0:50:25 ↗transcript
"I'm pretty excited about actually seeing change that improves human welfare a lot."
Baratunde Thurston said he worried about "a large-scale abdication", meaning AI could make us "super dependent and super weak." (Those words are the host's.) Kaplan didn't respond to the dependency point directly. Asked for his own vision, he gave the welfare and free-time picture quoted in the next section. Life with Machines · Dec 2024 · 1:11:44 ↗transcript
His positive vision, in four parts
1. AI that does physics better than we do
This is his most personal hope and the one he repeats most. It goes back to a childhood of science fiction and wanting to build a faster-than-light drive.
"I think that there's a very significant chance that we'll be able to build AI systems that are just better at theoretical physics than ourselves. And so I'm very excited to talk to those systems about what they think."
Asked whether anything could bring him back to physics research. IAIFI Symposium · Mar 2024 · 0:38:21 ↗transcript
"You can imagine AI systems working together to make the kind of progress that the theoretical physics community makes in, say, 50 years in a matter of days, weeks, et cetera."
This comes after "millions of AI models perhaps working together" doing "the kind of work that the entire scientific community currently does." Y Combinator AI Startup School · Jun 2025 · 0:10:44 ↗transcript
"There's no reason to imagine that… you can't have a creature or an AI system that is much more intelligent than we are and therefore would be much better at understanding the laws of physics than we are. And I do think that that is achievable."
He places humans on a spectrum running from "paramecium… mice… dogs… dolphins and us", with no reason for AI to stop at our level. Life Is a Story podcast · Nov 2020 · 0:54:02 ↗transcript
"I do think that intelligence, whether or not it's conscious or not, whether or not it understands or not, is a pretty profound and interesting thing about the universe."
He argues that AI should be seen as a scientific instrument like the Hubble telescope, not just "this kind of tool of revenue for big companies." NeurIPS 2021 panel · Dec 2021 · 1:19:19 ↗transcript
2. Cures, energy and human welfare
"My default guess is that AI is going to be incredibly beneficial. It's going to make our lives easier. It's going to give us access to better medicine, better ideas."
In the next breath he adds that "there's reason to be wary." Bloomberg Live · Oct 2023 · 0:14:53 ↗transcript
"Over time, we were able to use those development science and kind of modern life to cure diseases, allow us to have a lot more free time. Hopefully, I think better connections to people with that time, better ways to communicate."
His model is the Industrial Revolution: factory work was worse at first, then things got much better. He puts AI alongside "electricity, vaccines, modern science." Life with Machines · Dec 2024 · 1:12:22 ↗transcript
"Globally, there's still a lot of poverty, a lot of disease, a lot of problems that are solvable. And I think that we really have the possibility with big changes in technology of upgrading human welfare a lot."
"Maybe this technology can help us to cure cancer 10 times faster than we would otherwise."
He gives this as the benefits argument he and Dario make, in a passage about weighing it against "millions of geniuses in a data center." Azeem Azhar's Exponential View · Apr 2025 · 0:31:27 ↗transcript
3. AI that expands what people can do
"I think good AI is something that expands human agency, that makes it so that you can accomplish more of what you want to do."
He goes on: AI systems "should be kind and helpful and make us more productive. And they should basically kind of expand our ability to create a better world." Life with Machines · Dec 2024 · 0:20:55 ↗transcript
"The sort of optimistic vision is just that AI systems will be very empowering. People will be able to do more with less much more easily. You'll be able to build all sorts of spectacular things or learn spectacular things much more cheaply."
His answer to a 12-year-old asking how to future-proof a career. He followed it straight away with "I think it's really thorny." IAIFI Symposium · Mar 2024 · 0:50:13 ↗transcript
"I think AI is fundamentally very empowering and should be sort of broadly accessible."
He says it "shouldn't be up to us to completely dictate the value" and that "freedom to use the technology in the way that you want is really important." Azeem Azhar's Exponential View · Apr 2025 · 0:42:00 ↗transcript
"I think that we'll really see kind of the democratization of expertise. I think everyone can get very, very sophisticated advice."
4. Helping people understand each other
This idea comes up only once, but it's his only picture of AI changing how people relate to one another.
"Other people are an abstraction… But maybe AI can help us to, because the AI system doesn't have the limitations we have, it can read a lot of material, it can synthesize it. Maybe it can help us to understand each other better, work together better."
Asked whether he actually wants human-level AI. He starts with "Look, I grew up reading lots of science fiction… there are a lot of ways that the world can be better." Life with Machines · Dec 2024 · 1:06:20 ↗transcript
He doesn't describe what an ordinary day would look like, where people find meaning, or what he hopes for the next generation. The nearest he comes to a human role is that people will act "as managers" of AI work. Y Combinator AI Startup School · Jun 2025 · 0:19:07 ↗transcript
How it changed over time
- 2020–22 In physics seminars and technical talks, AI is a scientific puzzle on a predictable trend. The only vision is that a system smarter than us could understand physics better than we do. He already says "we kind of need to go slowly." NeurIPS 2021 panel · Dec 2021 · 1:00:37 ↗transcript
- 2023 After ChatGPT, he mostly sounds alarmed. He once hoped the future "would look like Star Trek or The Expanse", but now thinks it "might look like some other set of science fiction novels and movies." Stanford MLSys Seminar · Mar 2023 · 0:01:40 ↗transcript On a Johns Hopkins panel: "I think this is basically very scary. It's very uncharted territory." Johns Hopkins IDIES panel · Feb 2023 · 0:32:26 ↗transcript
- 2024 His most expansive year. He sets out an "optimistic vision" of empowerment, says he is excited to talk physics with AI, allows a hedged "utopian future", and gives his fullest picture on Life with Machines: cures, free time, less poverty, and people understanding each other better.
- 2025 His timelines get shorter and his risk language gets sharper: he confirms the "disempowered if there's a conflict" line, and talks about "AI aliens" that society might refuse. Azeem Azhar's Exponential View · Apr 2025 · 0:30:00 ↗transcript The positive material shrinks to short items: physics in days, cancer cures "10 times faster", democratized expertise. He says he worries about job impacts "in three years, five years." TIME100 Impact Dinner · Oct 2025 · 0:09:42 ↗transcript
The hedges are part of the vision
Kaplan rarely says anything hopeful without qualifying it in the same breath, and he is openly uncomfortable painting the picture at all.
- "Probably this is all going to be great and we'll be able to create, if it happens, we'll be able to create a utopian future. But I mean, maybe not." It's the only time he uses the word "utopian". ICTS colloquium · Apr 2024 · 0:51:48 ↗transcript
- "It's really hard to talk about this kind of stuff without it seeming like BS or like hype to sell the next technological product." Life with Machines · Dec 2024 · 1:13:50 ↗transcript
- "I'd like to believe in sort of this productivity bonus where everything will go well and people will just be more productive and it'll be great for everyone. But I think it is really uncharted territory." Johns Hopkins IDIES panel · Feb 2023 · 0:31:37 ↗transcript
- "So I think that there's a route for AI to be very capable and also very beneficial. And I think I want that. But I don't want the bad stuff." Life with Machines · Dec 2024 · 1:06:49 ↗transcript
- "Capitalism is very, very, very far from perfect in terms of making sure human welfare is improved, is broadly available." Azeem Azhar's Exponential View · Apr 2025 · 0:41:33 ↗transcript
The science fiction he actually cites
He says he "grew up reading lots of science fiction", and that his mother was a science fiction writer. Life with Machines · Dec 2024 · 0:04:30 ↗transcript He went into physics because "I wanted to figure out if we could build a faster than light drive." Y Combinator AI Startup School · Jun 2025 · 0:00:29 ↗transcript He names only a handful of works:
Some references in these recordings come from other people. Skynet and The Boys were raised by Baratunde Thurston, the "hamsters" and "silverback gorillas" image is Azeem Azhar's, and Bostrom's Superintelligence came up from Azhar too. In the Building Anthropic roundtable, the talk of vaccines, cancer cures, "100 AlphaFolds" and AI for democracy comes from other co-founders, not Kaplan.