As Vibe Coding Fades, Woz Offers Production-Ready Alternative
Ben Collins and Brad Eckert say they have been watching the app-building bubble deflate in real time.
“It’s been pretty dramatic,” said Collins, who is co-founder and CEO of new AI platform startup Woz. Since June, traffic to Lovable has dropped 40%. Bolt.new is down 27%. Vercel’s v0? Down 64% since May, all according to data from Barclays.
The two MIT-trained engineers, who were roommates over a decade ago, think they know why. Those tools were never meant to build real products. They were built to create demos.
“Everyone got really excited about going from a prompt to a working prototype in minutes,” Collins told The New Stack in a recent interview. “But then reality sets in. You’ve got this mess of code, security holes and no way to actually maintain it.”
So, they built San Francisco-based Woz, which launched last week with $6 million in funding to combine automation with actual human engineers to build apps that go beyond prototypes.
The Problem Nobody Wanted To Talk About
Ask any engineer who has worked at a big tech company, and they’ll tell you a similar story. A product manager uses one of these rapid prototyping tools, builds something cool and shows it to engineering. Engineering says, “Great, we understand what you want,” and then throws the whole thing in the trash, Collins joked.
“They start over from scratch,” he said. “Because the code is unusable. It’s what we call spaghetti code — everything is tangled together, you can’t tell what’s doing what and it’s full of shortcuts that will come back to bite you later.”
That works fine if you’re Google or Meta, and you have hundreds of engineers. But most businesses don’t.
“A restaurant chain doesn’t have an engineering team,” Collins says. “A consulting firm doesn’t have an engineering team. They need the final product, not a prototype they have to spend six months rebuilding.”
Injil Muhammad, co-founder and CEO at Buddle, a learning platform for Generation Z workers, tried the prototype route. “The quick tools are great for testing ideas,” he said in a statement. “But you hit a wall pretty fast when you need something that can actually scale and won’t break.”
Muhammad added, “Woz has been an absolute game-changer for our business. Vibe coding tools are great for getting ideas out there to be tested, but that next step toward sustainability requires technical expertise that is not easy to find. Woz’s team is responsive, incredibly easy to work with and asks the right kinds of questions to help you make your vision a technical reality. They’re real humans that are partners with you, and that makes all the difference.”
Assembly Line for Apps
Eckert, Woz’s CTO and co-founder, who previously founded a company that raised more than $40 million, had a realization that software development is not really that much different from manufacturing. You break down a complex process into steps, you standardize the components and you add quality checkpoints.
“We’re not letting the system reinvent the wheel every single time,” Eckert told The New Stack. “We’ve built these Lego blocks — payments, authentication, security — and they’re rock solid because my team has been building apps that scale to millions of users for over a decade.”
The team includes engineers who have sold companies to Yahoo and shipped products used by millions. “These are people I’ve worked with for ten years,” he said. “One of them was my boss at one point, then we switched roles. It’s not a junior team figuring things out as they go.”
The Woz system uses those prebuilt components as building blocks, combining them based on what the customer needs. When something complicated comes up, an actual human engineer steps in. And before anything goes live, a human reviews the whole thing.
“Every single app gets looked at by someone who knows what they’re doing,” Eckert says.
The Security Wake-up Call
Over the past six months, autogenerated apps have been getting hacked. Often, it’s because of shortcuts, Eckert said.
“It’s easier for an automated system to just dump user photos into a public storage bucket where anyone with the URL can access them,” he said. “We’ve seen that happen multiple times. Someone uploads their driver’s license photo, and suddenly it’s just sitting there on the internet.”
The Woz approach makes that technically impossible. Key security functions use those standardized Lego blocks. You can’t accidentally, or even intentionally, make files public, because that option does not exist.
“We don’t even give customers access to change security settings,” Collins said. “It’s locked down by default. You can’t mess it up.”
That matters more as they move toward bigger customers. Collins noted that Woz is talking to Fortune 50 retailers and medical device companies — organizations that spend tens of millions a year paying consultants to maintain clunky old systems.
“These companies aren’t playing around,” he said. “They need real security, real scalability. They can’t have something break because an automated system took a shortcut.”
The MIT Factory
There’s an old joke at MIT, Collins said: If you haven’t started a company by your sophomore year, you’re a failure.
“By that standard, we’re failures,” he laughed. “But I think we’re OK with that.”
Collins and Eckert met as undergrads, both studying engineering, both with startup ambitions. Then life happened. Eckert dropped out of his master’s program to join a tiny startup in Mountain View, Calif., with “fake grass on the floor, total garage situation,” he noted. That’s where he met the engineers who would eventually form Woz’s core team.
That company ran for seven years before Eckert was ready for something new. Collins had been building businesses, too. When the app-building boom kicked off last year, they both saw the same thing, they thought. “This technology was going to change everything, but not in the way people thought,” Colins said.
“Everyone was focused on making prototypes faster,” he explained. “We kept asking: What happens after the prototype? And nobody had a good answer.”
Their answer borrows from the Industrial Revolution. “The factory changed everything because it standardized production,” Collins argues. “That’s what we’re doing for software. Same quality, every time, at scale.”
Right now, that means mobile apps, but they’re planning to expand to web apps, TV apps, Internet of Things (IoT) software, even AR and VR. And heavily regulated industries like healthcare and insurance, where code quality isn’t optional.
It’s a new approach to product building, an AI App Factory. “Just as the physical factory powered the Industrial Revolution, Woz unlocks the next great leap in how the world builds,” Collins said in a statement.
What the Warriors See
For instance, Kirk Lacob‘s family owns the Golden State Warriors. They invested in Woz’s seed round, and they’re also customers.
“We’re not an engineering organization,” Lacob said. “But we need to put high-quality products in front of our fans. Woz makes that possible.”
There have not been great options for organizations like the Warriors that need apps but do not have engineering departments. Meanwhile, custom development is expensive and slow, and the DIY vibe coding tools create things that break, Collins said. Thus, Woz is betting there is a middle ground.
“We’re basically running a Shopify model,” he said. “We host it, we maintain it, we update it. You don’t need to hire engineers or figure out how servers work. You just tell us what you need, and we build it.”
Reality Check
However, Eckert is careful not to oversell what Woz does. “Look, we still think those rapid prototyping tools are great,” he said. “They’ve let people who could never code before see their ideas come to life. That’s genuinely amazing.”
The problem, though, is that people were sold a lie about what comes next, Eckert stated.
“They were told, ‘You can take this all the way to a real business with thousands of customers. That it’ll be secure and maintained and you can keep building on it,’” he told The New Stack. “That’s just not true. Not yet anyway.”
Woz is not looking to eliminate engineers from the development process. More specifically, it is looking to provide them as a service.
“You still need someone who understands what’s happening behind the scenes,” Eckert said. “That’s what we’re providing. Real engineers, at scale, at a price point that makes sense for businesses that couldn’t afford a full development team.”
It appears that people have figured out that building a prototype was the easy part.
However, “There’s a saying at MIT: Solve hard problems. Don’t pretend they’re easier than they are,” Collins said.
