Jun 22, 2026 · 1h 3m
Startups leverage live-map gamification and AI brain modeling to disrupt health tech
The hottest running app has nothing to do with speed | E2303
This episode highlights how non-traditional mechanics like gaming and advanced AI infrastructure are solving massive engagement and financial hurdles in health and biotechnology.
- 1INTVL drove over one million downloads by gamifying running through territory capture rather than focusing on athletic speed.
- 2Verge Labs pivoted from direct drug development to licensing AI infrastructure and massive brain datasets to pharmaceutical companies.
- 3AI-driven world models of disease can significantly lower the high failure rates and extreme costs of traditional drug discovery.
Don't miss
Alice Zhang explains how Verge Labs uses world models and transformer architectures to analyze brain tissue and slash drug development failure rates.
The brief
Louis Phillips built the running app INTVL by focusing on territory capture instead of speed. By turning daily runs into a live-map strategy game, the startup achieved over one million downloads and proved that gamification can drive healthy habits.
Rather than relying on expensive agency campaigns, Phillips attributes their growth to keeping marketing operations in-house and running highly targeted social media ads with a strict focus on lean customer acquisition costs.
In a sharp pivot to biotech, Alice Zhang explains how Verge Labs rebranded and shifted from internal drug development to building AI infrastructure. The company now licenses its massive brain dataset to help other pharmaceutical firms predict drug success.
Verge Labs utilizes advanced multimodal transformer architectures and world models of disease to analyze brain tissue and genetic data, aiming to solve the massive financial risk of high failure rates in traditional drug discovery.
What was said on this episode
33 statements · 24 positive · 4 negative · 5 neutral
Interval’s territory notifications increase users’ motivation to run.
“People are just so much more motivated to go out and do that activity.”
Listen at 0:08
Interval is a gamified running app built around claiming territory.
“Interval, which is a gamified running app”
Listen at 7:45
Interval intentionally does not currently score users by speed.
“right now there is nothing for speed inside interval, which was intentional”
Listen at 9:54
Interval enables runners of different abilities to compete against one another.
“it just brings in this element of anyone can compete against anyone”
Listen at 11:44
Interval plans arenas where users compete for the fastest daily performance at specific locations.
“what we're going to create and what is in works at the moment is something called arenas”
Listen at 13:30
Interval arena territory and rankings will reset daily.
“this will be resetting every single day”
Listen at 14:09
Interval reached approximately one million downloads and 100,000 Instagram followers.
“we grew that to about a million downloads and about 100,000 followers on Instagram”
Listen at 19:18
Interval’s Meta advertising cost per trial is approximately $12.
“the cost per trial starts for us currently is about $12 on Meta”
Listen at 23:35
Interval’s average customer lifetime is approximately 17 months.
“average customer lifetime is about 17 months”
Listen at 23:44
Converge is a target-discovery engine for identifying disease-causing proteins and designing drugs.
“what we built Converge originally is what we call, it's a, called a target discovery engine”
Listen at 31:34
Verge Labs has partnered with more than 24 global tissue banks, hospitals, and academic centers.
“we've partnered with more than 24 different tissue banks, hospitals, academic centers across the world”
Listen at 33:04
Verge Labs sequences approximately 30,000 genes across DNA, RNA, and protein levels.
“we sequence them so we capture the behavior of all 30,000 genes in the genome at multiple levels from the DNA to RNA to protein”
Listen at 33:32
Verge Labs can reconstruct a patient’s brain state from a single blood draw.
“create what we call a virtual biopsy of the brain”
Listen at 34:34
Transformer architectures can infer missing biological data across patient modalities.
“The power of these transformer based architectures is that it allows you to actually piece together missing data and infer missing data from what you have”
Listen at 35:43
Direct brain tissue is necessary to understand neurological disease mechanisms accurately.
“you need to go into the brain where it's happening”
Listen at 37:46
Brain tissue provides molecular ground truth for neurological disease models.
“brain tissue is like the lidar of neuroscience in that it's just the molecular ground truth of disease”
Listen at 38:13
Verge Labs represents each patient as a 512-dimensional vector.
“each patient is represented essentially as a 512 dimensional vector”
Listen at 39:54
Verge Labs’ multimodal architecture maps blood, brain, and genetic data into one shared space.
“Blood, brain and genetics can all live in the same mathematical language”
Listen at 40:40
Verge Labs’ model can reconstruct brain activity from blood alone with high accuracy.
“our own model, with high accuracy, can actually accurately reconstruct brain activity from blood alone”
Listen at 42:30
Verge Labs can identify likely responders and enable smaller, cheaper clinical trials.
“we can help you pick out which patients to roll in your next trial that actually respond to your drug and let you design a much smaller and cheaper clinical trial”
Listen at 43:42
Verge Labs’ therapeutic partnerships involve $25–42 million upfront and $700–800 million in milestones.
“it was a 25 to 42 million upfront with then milestones that total up to anywhere between 700 to $800 million each”
Listen at 44:23
Verge Labs says 83% of Lilly-partnership targets validated in wet-lab experiments.
“83% of those targets actually validated in wet lab experiments”
Listen at 46:20
Neuroscience currently trails cancer research by approximately 10–20 years.
“it's why I think neuroscience is long behind cancer by 10, 20 years”
Listen at 48:00
Biological AI gains have come more from scaling data and modalities than model parameters.
“the biggest gains we've seen have actually come from scaling data and modalities”
Listen at 49:07
Verge Labs plans to model patient disease progression longitudinally.
“we can actually create a virtual model of the patient in time where we can actually run forward”
Listen at 51:21
Early intervention in Alzheimer’s disease can alter disease progression.
“being able to intervene early enough to change your trajectory”
Listen at 52:12
Verge Labs observes nonlinear performance gains as it adds biological samples.
“we are seeing scaling laws in our data right where we're. And they're Nonlinear and increase as we add samples”
Listen at 53:10
Within five years, AI will support multiple stages of drug development.
“what I see in five years is really, I think AI will come into the pipeline at multiple points from multiple different models”
Listen at 55:05
AI will enable personalized medicine based on distinct disease subtypes and individual responses.
“that really brings us to a world of true personalized medicine”
Listen at 55:29
Developing one drug currently costs approximately $5 billion on average.
“it costs $5 billion on average all in to develop a single drug”
Listen at 56:31
Approximately 90% of drug-development attempts fail at the final stage.
“9 out of the 10 attempts fail at the last stage in the most expensive stage”
Listen at 56:46
Accurate drug-success prediction could reduce development costs from billions to tens of millions.
“if you can perfectly with accuracy predict kind of which drug will succeed, it goes from 5 billion to really tension tens of millions to really get a drug all the way through”
Listen at 57:12
A major risk is abandoning healthcare AI after an early setback.
“the biggest risk is more of a human one, which is that we kind of lose interest in AI or in the application of AI in healthcare just because we kind of face one step back”
Listen at 58:38
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions
This Week in Startups
Explore more interviews with cutting-edge tech founders and startup operators.