The 24-Hour Hackathon Where Non-Coders Beat Engineers

A CNBC correspondent with no technical background just helped build a functional app in 24 hours. That sentence would have been absurd three years ago. Now it's Tuesday.

Ernestine Siu spent her weekend at one of Singapore's largest in-person hackathons, where over 400 participants—from veteran engineers to complete novices—built software from scratch using AI coding assistants. What happened there tells us more about the future of software development than any venture capital thesis or corporate AI strategy deck.

According to CNBC's reporting published October 23, 2025, roughly half the participants had never attended a hackathon before. Some people who learned to "vibe code" just weeks earlier placed high in rankings, beating experienced engineers. The winner, Sritam Patnaik, built an AI-powered whiteboarding tool controlled by hand gestures tracked through webcam, taking home prizes worth approximately $39,000.

What Actually Happened

The event ran from Saturday morning through Sunday noon on a university campus, sponsored by OpenAI, Cursor, Anthropic, Google DeepMind, and other AI companies. Teams had one goal: build something functional using AI assistance.

Projects ranged from practical to absurd. F**Yu.AI is a productivity app that "bullies you into greatness" by calling users to yell at them about incomplete tasks. RizzedIn positions itself as LinkedIn for dating, connecting "career-minded individuals." The second-place winner created a human-versus-AI speed game. Third place went to a "Netflix for corporate training" that makes compliance videos engaging.

Siu's team—which she found through Discord—built Heirloom, a digital time capsule for preserving family stories and recipes across generations. Her contribution? "Ideas, storytelling abilities and vibes." Not code. That's the point.

An estimated 70 people worked through the entire night. Organizer Sherry Jiang, co-founder of fintech app Peek, described finding people "in random lecture rooms" at dawn: "It was like catching wild Pokemon."

The Barrier Dropped, The Bar Rose

"We've lowered the barrier, but raised the bar," Jiang told CNBC. That's the most concise summary of what AI coding tools have done to software development.

The technical barrier to building software has collapsed. People with "product sense and good taste" who understand positioning are now competing effectively against engineers with years of experience. Richard Lee, who built gamified habits app Orbie and slept 30 minutes on a lecture hall floor, said he felt he had "significantly upgraded" his skills in just 24 hours despite prior coding experience.

But the competition got fiercer. When anyone can build a functional prototype, differentiation comes from taste, positioning, and understanding what people actually want. The technical execution that once created moats is becoming commoditized.

What This Means for Teams and Companies

Lee and Jiang agree that both startup teams and corporate developer teams will shrink as AI-assisted coding tools improve. The math is straightforward: if three people can now accomplish what previously required ten, companies will adjust headcount accordingly.

Lee was direct about the stakes: "It's improving so fast that if you are not using these tools day to day, I think you'll be at a big risk of being eliminated."

That doesn't mean engineering skills are obsolete. Lee emphasized that understanding code logic and debugging remain essential. You still need to evaluate whether the AI-generated code actually works and understand how to fix it when it doesn't. But the nature of the work has fundamentally changed from writing most code manually to orchestrating AI tools and validating their output.

The Hackathon Scene Lost Its Soul

The event's organizers positioned this hackathon as an attempt to revive Singapore's builder community. Agrim Singh, co-founder of Niyam AI and co-organizer, wrote on LinkedIn that "the Singapore hackathon scene lost its soul," describing how events that once centered on "making something that worked" devolved into panels, sponsorship decks, and photo ops.

"Most AI events here feel hollow," Singh wrote. "Panels by people who've never touched the tech. 'Thought leadership' with no practical weight. People pretending to build, or worse, extracting revenue from the hype without caring about the ecosystem."

The organizers wanted to return to the original spirit: show up, build something real, ship it. The instruction was simple: "Use your imagination, build something unhinged. Build something wacky."

Apparently 400 people were ready for that challenge.

The Uncomfortable Implication

Here's what nobody wants to say directly: if a CNBC correspondent with no technical background can contribute meaningfully to building a functional app in 24 hours, what does that mean for people whose entire career value proposition is "I can write code"?

The answer isn't that those people become irrelevant. It's that their value needs to come from somewhere else—system design, architectural decisions, understanding business context, taste in what to build, ability to validate AI output. The pure mechanical skill of translating requirements into code is becoming less valuable as AI handles more of that translation automatically.

This is happening faster than most organizations are prepared for. Companies are still hiring like it's 2022, looking for engineers who can grind through tickets. The market is shifting toward people who can orchestrate AI tools, evaluate quality, and understand what should be built in the first place.

Time to Build Just Got Shorter

"The time to build is far shorter than before," Lee told CNBC. "It's really much easier for developers or technical people, or even non-technical people to build a prototype and effectively get to market."

That compression changes everything. Startup validation cycles accelerate. The gap between idea and prototype shrinks. The advantage of technical co-founders diminishes when non-technical founders can build initial versions themselves.

For marketing and growth leaders, this means product iteration speeds are about to increase dramatically. The old excuse of "that feature will take two sprints" starts sounding hollow when startups are shipping comparable functionality in hours. Customer expectations will adjust accordingly.

What Comes Next

Hackathons are useful signals because they reveal what's possible under pressure with current tools. What 400 people built in Singapore over 24 hours represents the new baseline for what small teams can accomplish with AI assistance.

The organizations that adapt fastest will be the ones that recognize this shift isn't about replacing engineers—it's about fundamentally restructuring how software gets built and who can build it. The ones that cling to old models will find themselves outpaced by competitors who figured out how to move faster with smaller teams.

And yes, that's uncomfortable. Most significant technological shifts are.

If you're trying to understand how AI tools change team structure, skill requirements, and competitive dynamics in your industry, our growth strategists can help you adapt before your competitors do. Let's talk about building strategies for a world where the barriers keep dropping.