School projects to startup dreams
Alex and Teddy met freshman year at Dartmouth, while both studying CS. Like most college friendships, it started casually, hanging out on campus, sitting next to each other in class, teaming up on group projects. Little did they know, one of those group projects would set the stage for them to build Magic Patterns.
It was called Left on Read and showed users analysis like their funniest message of the year and texting habits. Alex describes it as a “Spotify Wrapped for your text messages.” They shared the site on Reddit, and it blew up, going viral on campus. Alex says, “We were two kids in our dorm room, and people from all over the planet were analyzing their text messages with us.”
That experience gave them the first taste of what it felt like to build something on the internet, with thousands of people using it.
After graduation, they moved to San Francisco together. Teddy started working at Robinhood, and Alex joined LiveRamp. They were learning a lot while working for bigger companies, but they were also interested in all the startups emerging around them. Alex leaped first, becoming the first engineer at a YC startup and quickly bringing Teddy on to build with him. They had worked closely with the founders and had a front row seat to what it really takes to build a startup. They knew they wanted to build something together, not as employees, but as founders.
“It was always in the back of our minds,” Alex says about Teddy. “It wasn’t some big decision. It was just obvious. I’ve never once thought there was even an iota of risk in our co-founder relationship,” Alex says. “It's obvious having a bad co-founder relationship can kill the company. We have never ever worried about that.”
They decided to apply to YC and began tossing around ideas for their application, everything from waitlist software to a chess.com-style site. But in the end, they went with Left on Read, the one project they had built that already had real traction and users behind it.
The low that followed the YC high
Getting into YC was a rush. Alex and Teddy applied with Left on Read, knowing that it lacked the elements of a venture-backable business, but it was the clearest proof they could offer of what they could build together. The YC partners validated they’d been admitted not for their idea Left on Read, but for their potential as founders.
Instead of discouraging them, YC gave them the conviction that together, they had what it would take to figure it out.
They knew they needed to pivot, but they struggled to land on an idea that had legs. At one point, they were working on a healthcare tool. Teddy’s sister, who’s a doctor, gave them input, but after a week, they knew they were headed toward a dead end. “We had no idea what we were doing,” Alex says. “It was like, what are we even talking about?” Feeling lost in a space they didn’t know pushed them back to what they knew. With YC partner Dalton’s guidance, they refocused on frontend tools and design systems.
At one point, they shipped a Chrome extension that let developers “screenshot” a section of a live website and grab the underlying code. It was a kind of internal dev tool they described as screenshots for engineers.
They leaned into it for a minute, even pitched it to their YC partners.
“We told them, we're 'building screenshots for engineers',” Alex recalled. “And they said it back to us in this way that made us immediately realize—wait, this is not what we want to be: 'screenshot guys'”
Demo Day kept creeping closer. Most companies see it as a launchpad, but for Alex and Teddy, it felt like a deadline where they’d have nothing to show. They had only one paying customer for the component editor tool they’d been working on that paid $50/month, which meant they only had $600 ARR. They felt embarrassed as others in their batch reported $100K+ ARR.
It went as badly as they expected. They struggled to get interested investors and didn’t end up raising much money. But they still had cash in the bank and deep faith in themselves to continuing to work together, especially in design tooling.
Rebuilding with better AI
Months after Demo Day, they decided to revisit the latest LLM capabilities to see if things had improved and might open new doors for them. Turns out it did, and much better than they had expected. So they revisited an early prompt-based website builder idea and turned it into a fun hackathon, setting aside two days to work separately on their own versions and see what would come out of it.
Teddy built a public component gallery, essentially a way to browse the full spectrum of React building blocks across the web. Alex focused on something more generative, using natural language to create React components from scratch. When they regrouped, the two projects clicked into place.
That merge became the foundation of Magic Patterns, an AI-first design tool that lets users build, remix, and design with simple prompts. Everything they’d been circling for so long, design tooling, AI, frontend dev, suddenly came together in one focused product.
And surprisingly, "screenshots for engineers" reappeared. It was perfect for letting users grab live components from any website and instantly repurpose them with AI.
Almost a year after getting into YC, and after months of pivoting through ideas that didn’t feel right, they launched Magic Patterns publicly in September 2023.
But that same month, Vercel launched v0.
“People told us, you guys are cooked. That was basically the sentiment,” but Alex and Teddy were more motivated than ever. They took it as a hugely positive sign that we were onto something big. “If they’re investing here, it means the space is real.” They ignored the noise and kept going. “We were more excited than ever! We had customers and users!” says Alex.
Doing the work of 100 with 2
Alex and Teddy didn’t realize their first real inflection moment was happening until afterward. A user from a company in Australia called Lendi Group signed up on their own. A few days later, Lendi invited a teammate to a paid seat. Then another. Alex reached out to schedule a call, offering to meet at whatever time, and before he could speak to them, they’d added even more paid seats. “I was legit looking at flights to Sydney to potentially meet and onboard the company in person, and then they just signed up self-serve.” Now, he and Teddy were staring at a signal they knew they had to figure out how to repeat.
At this point, Magic Patterns is just Alex and Teddy. Their average customer support response time was eight minutes, even when they had more than 1,500 customers. If a customer churns, Alex emails them directly with an apology. “If someone isn’t successful with the product, that’s on us.”
From this point forward, they just kept talking to users and shipping features. Under the hood, they were consistently growing double-digits every month and became profitable.
Customers started to share publicly how much they love the product and appreciate the timely support. Soon Lenny called it one of his favorite AI tools for vibe-coding; Claire Vo featured it on her How I AI podcast with Colin Matthews; and mentions continued to grow.
Alex says, “I think that speaks to loving the customer and caring deeply about building a really good product. I'm so glad we're creating value for so many people. And of course, it feels amazing that we are mentioned in the same breath as the other bigger players. Much, much, much bigger companies than us are being mentioned alongside Magic Patterns.”
It came from steady, heads-down execution. Just Alex and Teddy, building fast, listening to users, and pushing the product forward while the momentum quietly compounded behind them.
Building the future of software development
Growth hasn’t just picked up speed, it’s taken off like a rocket ship. Alex and Teddy are chasing a future so ambitious, it can’t be built alone. It’s the kind of vision that demands new infrastructure, new interfaces, and builders bold enough to help reinvent how software gets made.
Their long-term vision isn’t to build another design tool. Alex and Teddy see a future where the way software gets built looks fundamentally different. In that world, Magic Patterns isn’t just a tool, it’s the foundation everything runs on.
“We want to be the one-stop shop for frontend in an AI-first world. People are no longer dragging rectangles to make mockups and prototypes”, Alex says. “The product manager and the designer will be the folks creating software. And these are very, very early days. That’s the vision we’re building toward.”