The good future, according to Ben Mann
What Mann says a good outcome with AI looks like, drawn from the eight talks, panels and podcasts of his that we could find and transcribe. There isn't much material. Most of his vision comes from two 2025 interviews and one off-script answer on a 2023 panel. The page also covers whether he ever talks about Iain M. Banks's Culture, or about humans ending up as AI's pets.
The short answer
Banks, the Culture and benevolent "Minds" appear in none of the eight appearances. The nearest he gets is the science fiction he grew up on: "space operas where humanity is a multi-galactic civilization, has extremely advanced technology, building Dyson spheres around the sun with sentient robots to help them." Lenny's Podcast · Jul 2025 · 0:34:52 ↗transcript That sounds a bit like the Culture, but he names no author or book.
Neither Mann nor any host uses pet, zoo or ward imagery. In his version of the future, AI serves what people really want, and people manage large teams of AIs. The one "AI overlords" remark in these recordings is the host Lenny Rachitsky joking, and Mann doesn't pick it up. His only picture of humans being sidelined is the catastrophe in which a misaligned AI "will like not even worry about the humans because it's mostly focused on getting more compute." Berkeley LLM Agents MOOC · Nov 2024 · 0:19:24 ↗transcript
Close to the question, but not quite
He touches on a few of the questions around the pets idea: what people would do with free labour, how far to hand control to AI, and whether humans are meant to be idle.
"In a world of abundance where labor is almost free and anything you want to do, you can just ask an expert to do it for you, then what do jobs even look like?"
He asks the question himself, then declines to answer it: "hopefully we'll have figured it out. And in a world of abundance, maybe the jobs themselves, it's not that scary." Lenny's Podcast · Jul 2025 · 0:13:53 ↗transcript
"Understand us better than we understand ourselves… do what we meant, or meant if we were 10 times smarter and had 10 years to think about it"
This is the nearest he comes to a benevolent-guardian AI. He frames it as AI serving people's own better judgement, not as people being looked after. SaaStr panel · Jan 2024 · 0:27:09 ↗transcript
"Superintelligence is a lot about like, how do we keep God in a box and not let the God out?"
He says it is "both hilarious and terrifying" to watch people hand language models the internet and their bank accounts. Lenny's Podcast · Jul 2025 · 0:36:45 ↗transcript
"I'm not immune to job replacement either. So just some vulnerability there of like at some point it's coming for all of us."
"It's really unlikely that we evolved to just be at leisure."
This comes from Nate Soares's idea of "resting in motion", and he says it about his own stress, not about AI. Still, it cuts against the picture of humans retired into leisure. Lenny's Podcast · Jul 2025 · 1:01:08 ↗transcript
His positive vision, in five parts
1. AI that does what we meant
His most personal hope came on a 2023 SaaStr panel, when the moderator asked each speaker for "something wild" they hoped would be true by 2030. It is close to the idea of AI acting on what we would want if we were wiser.
"I guess my big bright future is that language models understand us better than we understand ourselves. And when we ask it to do things for us, it will not just do what we say, but do what we meant, or meant if we were 10 times smarter and had 10 years to think about it, that's what it would do."
"And it would make us all better people and improve our relationships and help us be the best versions of ourselves. That's my big bright future."
He added that the realistic outcome might fall short: "On OKRs, they say 60% is success. So I take that." SaaStr panel · Jan 2024 · 0:27:36 ↗transcript
"We want the AI to be like, oh, obviously what you really meant was this, and that's what I'm going to help you with."
The same idea two years later. He contrasts it with the monkey's paw and King Midas, and says this is why Claude's personality is tied to safety. Lenny's Podcast · Jul 2025 · 0:30:18 ↗transcript
2. A country of geniuses, and science that tests itself against reality
By 2025 he uses Dario Amodei's Machines of Loving Grace language directly. For him, the way AI gets beyond human expertise is empirical: running experiments and learning from the results.
"If we do our jobs right, we will have safe, aligned superintelligence. We'll have, as Dario says in Machines of Loving Grace, a country of geniuses in a data center. And the ability to accelerate positive change in science, technology, education, mathematics, it's going to be amazing."
"If you want to spin up six million cancer researchers overnight to go cure cancer, you can totally do that."
This is what he thinks becomes possible once AI passes his "economic Turing test", where hiring managers can't tell an AI contractor from a human one. JoCoding · Nov 2025 · 0:27:52 ↗transcript
"We could expect them to sort of improve recursively in an environment and potentially become much better than humans could be just by banging their head against reality"
"It used to take them like 12 weeks or something to write a report on cancer patient… And now it takes like 10 minutes to get the report."
His concrete example, from Anthropic's work with Novo Nordisk. He says the results then get measured and fed back into the system. No Priors · Jun 2025 · 0:24:06 ↗transcript
3. Abundance after the singularity, which he won't describe
He expects the economy itself to change beyond recognition. But each time he gets near describing everyday life, he says it can't be forecast.
"Where we're way past the singularity, it's hard for me to imagine that even capitalism will look at all like it looks today."
This is about 20 years from now. Lenny's Podcast · Jul 2025 · 0:13:22 ↗transcript
"In the immediate term, there will be a massive expansion of the pie in the amount of labor that people can do."
In the same answer he adds that "there will be a lot of displacement" in lower-skill jobs. Lenny's Podcast · Jul 2025 · 0:17:15 ↗transcript
"If the amount of goods and services in the world is doubling every year, what does that even mean for me as a person living in California, let alone somebody living in some other part of the world that might be much worse off?"
4. People as creators and managers of AI teams
"Maybe in three years every developer will be like the CEO of their own company in some way and they have a whole 2,000 person team of Claude codes"
He means the next one to three years, "before we get to superintelligence." JoCoding · Nov 2025 · 0:13:42 ↗transcript
"That allows people to create software products who aren't technical at all. That's a really big change."
"MCP, I think, is sort of a democratizing force"
This is about open standards spreading AI capability widely. No Priors · Jun 2025 · 0:39:42 ↗transcript
5. Human flourishing, decided together
"How do we ensure the continued… Flourishing of humanity into the indefinite future. That's the question I'd love to know."
His "real" answer when asked what he would ask an AGI, after two joke answers. Lenny's Podcast · Jul 2025 · 1:08:29 ↗transcript
"This is also not something that we think as a small group of people in San Francisco should be figuring out. This should be a society-wide conversation."
This is about the values written into Claude's constitution, and Anthropic's work on a "collective constitution". Lenny's Podcast · Jul 2025 · 0:32:40 ↗transcript
"As AI gets more and more powerful, we'll cede more and more control of Anthropic to elected board members who have no financial interest in Anthropic"
This describes the Long-Term Benefit Trust, which is his answer to the worry about power concentrating. Berkeley LLM Agents MOOC · Nov 2024 · 0:43:35 ↗transcript
"I guess I would say the future is sort of all of ours to create together. And this technology carries tremendous promise."
"At this point, I don't think any of it's going to matter. I just want her to be happy and thoughtful and curious and kind."
This is about his daughter. He says he would once have steered her towards a top school and lots of extracurriculars, and now he values her Montessori school's focus on curiosity. Lenny's Podcast · Jul 2025 · 0:22:49 ↗transcript
How it changed over time
- 2023 His conference talks are scripted and focused on safety. The positive future shows up only as a closing line: "charting a course toward a flourishing and just AI-powered future." The AI Conference · Sep 2023 · 0:14:13 ↗transcript When asked off-script on the SaaStr panel, he gives his most personal vision: AI that understands us better than we do and makes us better people.
- 2024 At Manifest and Berkeley he talks about the Responsible Scaling Policy, and the vivid scenarios are all catastrophes: Clippy breaking out of its box, and a planet turned into computronium. The good future is described only as the opposite of these: "a future with AI that benefits us all." Manifest · Jun 2024 · 0:24:17 ↗transcript
- 2025 His timelines firm up. He cites AI 2027 and a median of around 2028 for transformative AI, and he adopts Machines of Loving Grace language: a country of geniuses, curing cancer, abundance, capitalism transformed. He says more than once that his risk estimate has fallen because language models turned out to understand human values: "my estimation of the risk has actually gone down quite a bit." JoCoding · Nov 2025 · 0:35:26 ↗transcript
The hedges
He is optimistic overall, and he always puts caution first.
- "I think my personal feeling about this is that things are overwhelmingly likely to go well." Lenny's Podcast · Jul 2025 · 0:42:16 ↗transcript
- "I think my best granularity of forecast for could we have an X-risk or extremely bad outcome from AI is somewhere between 0% and 10%." Lenny's Podcast · Jul 2025 · 0:51:13 ↗transcript
- "Yes, we want to create safe AGI and deliver the benefits to humanity. But let's make triple sure that it's going to go well." Lenny's Podcast · Jul 2025 · 0:43:13 ↗transcript
- "A point beyond which you can't easily forecast what's going to happen. It's just such a fast rate of change and so different that it's hard to even imagine." Lenny's Podcast · Jul 2025 · 0:14:22 ↗transcript
- "Social structures and society and even the laws in different political areas, those will affect the speed of the rollout" JoCoding · Nov 2025 · 0:28:10 ↗transcript
- "It's going to be much weirder very soon." Lenny's Podcast · Jul 2025 · 1:09:27 ↗transcript
The science fiction he actually cites
He says reading science fiction growing up "positioned me to think about things in a long-term view", and that Bostrom's Superintelligence later made the risk feel real. Lenny's Podcast · Jul 2025 · 0:34:52 ↗transcript The works he names are:
He credits "the future is already here. It's just not evenly distributed" to Arthur C. Clarke. It is usually credited to William Gibson. The "AI overlords" line on Lenny's Podcast is the host's.