Bryn, welcome. Thanks for speaking with us. Tell us what you're working on.
Pleasure to be here. I'm building Flow Learn, an AI-driven SAT prep platform designed to personalize learning, making study more efficient and affordable for students.
Flow Learn adapts continuously to student needs, offering personalized sessions and real-time feedback. It's 30 times more affordable than traditional tutoring, available 24/7.
This isn't your first category-creating product, is it? Tell us about one or two others.
Before Flow Learn, I was head of product at Globe, a new trading exchange backed by YC. I joined as the first non-developer in a team of brilliant founders and scientists. When I left three years later, we had facilitated $15B in transactions and built what might be the most performant exchange in history.
Before that, I was Head of Product at Distributed, where I created a 'managed marketplace' for remote teams and a new category we called "Elastic Teams." Clients could onboard high-value teams—developers, data scientists, designers—within hours, rather than weeks, through our platform.
How does your process differ between building incremental and transformative products?
'Transformative' sounds more glamorous, but I think both incremental and transformative work are crucial, and even transformative products consist of many incremental improvements.
Take SpaceX. Most of its changes were incremental, but certain features, such as its engine and re-usability, are transformative.
Product development, whether incremental or transformative, starts with identifying key areas to target, validating ideas with customers, and iterating based on feedback. Incremental work focuses on feature adjustments, benchmarking against competitors, and responding to data. With transformative innovation, it's harder to predict usefulness or find competitors to benchmark against, so you need to embrace uncertainty. This is where conventional product strategies often fall short, leading some experts to dismiss truly transformative ideas.
Bringing a rigorous process to these challenges is something I'm passionate about.
How do you approach 'transformative' features or products?
There are a few frameworks I find essential for building transformative products:
The first is the first-principles approach, as popularized by Elon Musk. Instead of improving existing models, break things down to their basic components and rebuild. It's slower but can lead to revolutionary outcomes.
The second is the concept of causal cakes, which I learned from a philosopher of science. In short, outcomes often depend on multiple factors working together. For example, reducing class sizes improved results in Tennessee, but didn't replicate elsewhere because the surrounding conditions weren't right. Data alone can be misleading if you don't account for these interdependencies.
The third is local vs. global maxima. Sometimes, we focus on optimizing small improvements (local maxima) but fail to realize there's a bigger opportunity nearby (global maxima). You have to step back regularly and reassess if you're on the right path.
EdTech has been a tough market for startups. Why are you confident now is the right time to enter?
It's a massive market—over $5 trillion annually—but much of it flows to public education, with slow sales cycles and high demands for proof. Historically, EdTech hasn't made more than incremental improvements in K-12 or test prep.
As someone who's used, analyzed, and built EdTech products, I see too much friction for both learners and teachers. Great products still require time-consuming adoption processes, which busy users can't afford.
AI changes this. Conversational apps are intuitive and frictionless, and AI can provide instant, tailored feedback and generate content on the spot. This is the missing piece we've been waiting for, and I'm confident Flow Learn will be a key player in transforming the space.
