Esther Dyson was in an Uber when she joined the StartUp Health community for this conversation. She was heading to Detroit airport, fresh from a Boys & Girls Club awards ceremony in Lansing, MI, where she'd watched young people from underserved communities be recognized for the courage it took to make something of themselves. "It was the human beings that made them have the courage to step up," she noted. "Not AIs or devices or anything."

That observation, made almost in passing, became a thread running through the entire conversation that followed.

Dyson, investor, thinker, and longtime StartUp Health Impact Board member, has spent decades at the intersection of technology, health, and social equity. Her forthcoming book, Term Limits: Human Time and AI Scale, due in 2027, is the culmination of years of questions she has been trying to answer, not a manifesto of conclusions she already holds. "I didn't start this because I felt I knew everything," she explained. "It was more like, I have all these questions I want to understand."

The core premise of Dyson's book is deceptively simple: human beings are finite, and that finitude is not a problem to be solved. It is a feature. It gives us permission to prioritize, to share, to pass things along. "Don't delude yourself that you're the only one that matters," she said. "You want to make a dent in the universe. You don't want to control everything." She described the recognition of limits as calming, even liberating. "It's okay to do what you can. Figure out what you can do best, and then find other people, help them do what they can do best."

The conversation quickly turned to what happens when that logic collides with AI's seemingly limitless scale. Dyson described two diseases of time that she explores in the book: addiction, which is the compulsion for needing more and needing it right now, and an equally dangerous yearning for having more for forever. Both lead to the same place, an inability to choose, to trade off, to be satisfied with enough. "He knew the price of everything and the value of nothing," she said, quoting Oscar Wilde. That distinction between price and value runs throughout her thinking.

For health innovators, the implications are direct. Community member Matthew Swanson of Clutch Health raised the tension many builders face: using AI to scale patient engagement while still preserving the humanity that makes that engagement matter. Dyson pointed to two investments she finds instructive. The first is a peer-to-peer counseling platform where AI supports moderators, flagging when someone in a chat may need more urgent clinical intervention while the moderators themselves focus entirely on the humans in front of them. The second is a platform that trains neurodivergent people on workplace skills, using AI to handle the administrative and credentialing burden that otherwise eats into counselor time. In both cases, the ratio matters. "One person to four people is great. To talk to moderators, just one AI does it all." The AI extends human capacity; it does not replace human presence.

"Use AI to save humans time so they have more time to be human," she said. "AI should be relieving us of stupid friction so that we can experience the good friction." The good friction, she clarified, is holding somebody's hand, delivering words of comfort. "It may not be efficient, but it's what's needed at the moment."

Emmett Schaller of HELD, a brand new member of the StartUp Health community, pressed on a different challenge: the risk of building health products that become addictive by design, even with good intentions. Dyson's response was characteristic of her broader argument. "Consuming versus experiencing" is the real red line. She described her own experience after cancer of the esophagus, monitoring her sleep score on an Oura ring, trying not to eat after 4:00 PM, and the pull toward optimization that can itself become an obstacle to living. "My goal is not to be the healthiest person in the world. My goal is to have enough health so that I can enjoy my life and be productive." Everything in moderation, she argued, is a more durable principle than too much ain't enough. There is even a clinical term for going too far in the other direction: orthorexia.

She was generous with her caution, though, and equally generous with her acknowledgment of luck. "Most of us here on this call are in the healthcare business, doing something commercially successful. We have all been lucky." She drew a sharp contrast with the communities she has spent a decade alongside through Wellville. For people without financial resilience, taking a risk does not mean losing a contest. It means losing a home.

Community member Piyush Gupta of Ambrosia pushed the conversation into structural territory, asking why the system consistently incentivizes ongoing treatment over actual cures. Dyson's answer was pointed: because dependency is more lucrative. "The best forms of health are the ones where you trigger the body to be healthy versus you make the body dependent on a drug." She used GLP-1s as a live example of the question playing out in real time: will they reset the body like a vaccine, or require indefinite use like insulin? The answer has moral weight, not just clinical weight. "That's why we want human beings in healthcare, not bots that are trying to optimize profits."

She stopped short of full cynicism, noting that a genuine cure would also be a successful business if someone could pull it off. The deeper problem, she argued, is not within the healthcare system itself but in the broader public investment architecture around it. Savva Kerdemelidis of Public Good Pharma, joining from a biotech hub in Honduras, pressed her on the strongest policy fix for a sick care model built around short-term profits. Her answer was direct: single-payer healthcare, funded as public infrastructure, the same way society funds roads and bridges. "You can't expect poor people to be able to fund the flourishing of their own children. But having those children flourish in the future is of public value. It's a public resource and a public asset."

From there she would redirect resources toward early childhood education, doulas, better food, and taxes on sugar and ultra-processed food. The through line across both exchanges was the same: we keep confusing value with valuation, and paying for the wrong one. "We don't seem to be investing much in the flourishing of the human race," she said, "and somehow we don't seem to see that as a problem."

The signals she picks up traveling confirm it. Across communities, in poor places especially, what people want most is not a better app or a smarter device. "The people in poor places want dignity and respect, and they need a sense of agency." The children need more time with healthier parents, she added, and when that isn't possible, society needs to step in. The Boys & Girls Club ceremony she had just come from was a living example of what that looks like when it works.

The conversation closed with Dyson's most expansive and honest answer. Asked about the long-term trajectory of AI and its outer limits by longtime community member Bentley Adams of Way, she was candid about the limits of her own foresight. She is not worried about superintelligence as science fiction describes it. She is worried about the next five to ten years, when the choices humans make about how to deploy these tools are still very much within human control. "The real question is, are we going to be wise enough to use this stuff effectively, or are the smart people going to take it away and just be so good at creating predatory AI that we really get into trouble?" Wisdom, she said, is not the same as intelligence, and no amount of compute produces more of it. "Some guy said there's going to be a million Einsteins. That's not it. It's not being smart, it's being wise."

She will be at StartUp Health's Civic Health Forum in New York on June 16, continuing the conversation in person. The book arrives in 2027.