TL;DR Max Marchione, founder of Superpower, believes healthcare should become like the iPhone: high quality, low cost and available to billions. His approach is to picture healthcare 30 years from now and start building it today, using three systems (record, intelligence, action) and a set of product principles that make people want to keep using the product. The reason to start now: the longer we wait, the more people suffer.
At NimbleFest 2026, Max shared his manifesto for AI-native healthcare. Superpower is a preventive health company that makes biomarker testing accessible at a fraction of the traditional cost. Hundreds of thousands of members use it, and many of them describe the outcome as life-changing, some as life-saving.
His thesis fits in one sentence: healthcare should work like the iPhone. High quality, low cost, in the hands of billions.
Why start from healthcare 30 years from now?
Max's first frame for thinking about the future is counterintuitive. Predicting the next two years is hard. Predicting the next 30 is easier.
In 30 years, healthcare will be more preventive, higher quality and more tech-forward, and consumers will have more control. Picture an AI agent with full medical knowledge and your complete history, connecting you directly to diagnostics, therapeutics and referrals.
Then comes the question that drives Superpower: why wait 30 years? The technology to start already exists.
What keeps its value in healthcare after AGI?
Four categories keep their value once AI gets very good: infrastructure, intelligence, brand and IP.
Infrastructure. Physical assets like labs, manufacturing, compounding and biologic synthesis. As AI agents deal with manufacturers directly, consumer marketing margins go to zero.
Intelligence. In the medium to long term, you either build foundational models or own proprietary data. In the short term, good memory orchestration and multi-agent application layers still create value.
Brand. People still want to signal who they are, even when products get cheap. Max pointed to the Oura Ring.
IP. Diagnostic and therapeutic IP stays critical. Max sees therapeutics as the most important health category of the future.
What are the three systems of an AI-native product?
An AI-native product needs a system of record, a system of intelligence and a system of action.
A system of record captures and structures data: biological data, patient history, wearables, genomics and personal preferences.
A system of intelligence runs inference on that data and connects the dots. It also takes in medical publications and proprietary datasets. The key ingredient is the preference function: knowing someone's budget, how many recommendations they can take and how they live makes care personal.
A system of action turns insight into something that happens in the real world. Ordering a test, sending a nurse, writing a prescription, making a referral. The interface itself becomes a tool the AI controls.
The same lens works for a company like Salesforce. CRM-style tools that lack these three systems risk becoming obsolete.
What are Superpower's product principles for AI-native healthcare?
Seven principles shape how Superpower builds products people keep using.
Dynamic UIs. Interfaces that change based on your data and on what the AI recommends, instead of staying static.
Hero interfaces. A distinctive visual signature people recognize instantly, the way they recognize the iPhone home screen.
Personalized preferences. Onboard people with questions, feed their answers into a preference function, and keep updating it as the AI learns.
Delete. Cut features, screens, code and input tokens. A few things done well beat many things done poorly, and fewer tokens also makes AI inference perform better.
Explain. Show the reasoning behind a recommendation and let people expand it to see how the AI got there.
Support. Health journeys can be hard, and the AI should be emotionally sensitive to that. For one member with chronic back pain, the AI found treatment options covered by insurance. That challenges the assumption that empathy needs a human doctor.
Engage. Good products reach out first. Superpower sends AI-generated emails and texts that follow up on biomarker results before people start to worry. In one example, a member had a high CRP value. The AI pointed to recent exercise as a likely cause and suggested a retest.
Why build AI-native healthcare now?
It all comes back to the people waiting for better care. The longer we wait, the more people suffer. Health done right is what lets people create, make progress and live fully.
What to take home
You don't need to work in healthcare to use Max's frames. Picture your industry 30 years from now and ask what's stopping you from building part of it today. Check whether your product has a system of record, intelligence and action, or just one of the three. And look at what you could delete.
