AGEIST Using AI to Predict Death with Sune Lehmann
In this week’s episode, we welcome Sune Lehmann, a network and complexity scientist at the Technical University of Denmark. Sune shares insights into his revolutionary work using algorithms to predict health outcomes, income, and mortality, drawing on extensive datasets from Danish citizens. He emphasizes the power of viewing human lives as sequences of events, akin to language, and how this perspective can lead to groundbreaking predictions about individuals' futures. We also get into the ethical considerations and future implications of this research, highlighting the balance between predictive accuracy and the importance of shared risk in insurance. Additionally, he touches on the once-Danish-now-global phenomenon of cold water swimming, illustrating how this practice offers a unique sense of well-being and mental clarity.
What kind of SuperAger are you? Check the SuperAge Quiz and find out! (visit: ageist.com/quiz)
Thanks to our sponsors:
LMNT Electrolytes — our favorite electrolytes for optimal hydration. Listeners receive a free 8-serving sample pack with their purchase at DrinkLMNT.com/AGEIST
Timeline Nutrition — our favorite supplement for cell support and mitochondrial function. Listeners receive 10% off your first order of Mitopure with code AGEIST at TimelineNutrition.com/ageist
InsideTracker – the dashboard to your Inner Health. Listeners get 20% off on all products at InsideTracker.com/AGEIST
Key Moments
“From a certain perspective, human lives are sequences, right? You're born, you are seen by a doctor. They do some measurements and all of that information can be structured to create, let's say, the sentence of your life.
And so what we did is that we took an enormous dataset and we used the same technology that lies underneath the AI large language models, and we built a kind of general model for human lives.”
“That model, as long as you have training data, can really predict anything about humans. I mean, it might do things with kind of varying quality, but for the things that we've tested it, it performs better than any other algorithm that we can find.”
“So lots of people have kind of contacted me saying let's let's use this for something that's actually meaningful and useful. And I think, you know, what's amazing about what we do is that we also model not just the medical data, but the data around the life that people live.”
Connect with Sune