Interval intentionally excludes speed as a competitive factor. Users claim territory by completing a running loop, allowing anyone, regardless of pace, to compete. This broadens the user base from elite athletes to casual walkers, like a "grandma next door."
Interval's core loop is driven by notifications that a user's territory has been stolen. Attaching a specific person's name and face to the action makes the competition feel personal, creating a powerful motivation for users to go out and reclaim their turf.
Interval's founder grew the app to 1M downloads by creating simple, talking-head style explanation videos. He attributes this success to a willingness to fail publicly and not worry about looking like an idiot, a hurdle many founders struggle to overcome.
While organic social media drove Interval's initial growth, it produced volatile results with "astronomical" highs and "flop" lows. Implementing paid ads on Meta created a predictable user acquisition funnel, smoothing out growth and de-risking the business.
Verge Labs warns that the biggest risk to AI in healthcare is losing interest after early failures. True breakthroughs come from iterating and learning. Instead of burying a failed trial, they published the results and used the data to improve their platform, viewing it as a "hard-won lesson."
Verge Labs initially focused on discovering its own drugs. The experience taught them a more valuable problem is predicting which patients will respond to a specific drug. They pivoted from trying to win the lottery to selling "a better machine that sells those lottery tickets."
A key challenge in neurology is that brain data comes from deceased patients while trials use living ones. Verge Labs uses transformer AI to bridge this gap, inferring missing information and fusing disparate data sources (brain, blood, clinical records) into a unified "virtual biopsy."
Interval's founder clearly articulates his unit economics for paid acquisition. A $12 cost per trial start is justified because the average customer lifetime is 17 months. At approximately $60/year, this yields an LTV around $85-$90, demonstrating a healthy return on ad spend.
Verge Labs' key defensibility is not its AI models, which can be replicated, but its unique dataset of over 12,000 human brains from deceased patients. This decade-long, logistically complex data collection effort creates a powerful, hard-to-replicate competitive advantage.
The $5 billion cost to develop a drug is primarily driven by the high failure rate (9 out of 10) in late-stage trials. AI's biggest financial impact will be predicting which drugs will succeed, drastically reducing wasted R&D. This efficiency is what will ultimately make drugs more affordable.
Unlike text-based LLMs where simply increasing parameter count works, Verge Labs found the biggest AI performance gains in biology come from scaling data modalities—adding new types of data like proteomics and imaging. Fusing different data sources is more critical than just making the model bigger.