Two Journalists Built a $5B App Clone in One Hour Using AI

Two journalists with zero coding experience just replicated the core features of a $5 billion software company in under an hour. The cost? About $15 in AI compute credits. If that doesn't make software investors nervous, it should.

The One-Hour Experiment That Changed Everything

CNBC reporters Deidre Bosa and Jon Fortt set out to test a simple question: How real is the AI threat to software companies? Their weapon of choice was Anthropic's "Claude Code," an AI tool that builds functioning apps based on plain English commands.

Their target: Monday.com, the project management platform worth $5 billion.

The results were immediate. Within minutes of asking Claude to "build a project management dashboard similar to Monday," they had a working prototype complete with multiple project boards, team member assignments, and status dropdowns.

But the real magic happened next. The AI researched Monday.com independently, identified key features, and began adding them autonomously. A calendar appeared. Email integration followed. Soon, their clone was functioning as a personalized project manager, surfacing forgotten calendar invites and adding travel booking reminders.

Total time invested: less than 60 minutes. Total cost: $5 to $15 in compute credits.

The Exposed and the Protected

Silicon Valley insiders aren't surprised by these results. They've been quietly categorizing software companies into two buckets: the vulnerable and the defensible.

The most exposed? Companies that "sit on top of the work" — tools like Atlassian, Adobe, HubSpot, Zendesk, and Smartsheet. These aren't core business infrastructure; they're convenience layers that can apparently be replicated with a weekend coding session.

The safer bets occupy different territory. Cybersecurity companies like CrowdStrike and Palo Alto Networks benefit from network effects that no one wants to replicate and maintain independently. Systems of record like Salesforce anchor themselves with enterprise data, making them harder to clone with casual AI experimentation.

But even "safe" isn't immune. The question isn't whether AI can replicate these tools — it's how quickly and how cheaply.

The Investment Calculus

This year's wholesale sell-off in software stocks suddenly makes more sense. Investors are beginning to separate the "need-to-haves" from the "nice-to-haves," and AI is accelerating that process.

Consider the math: If a $15 AI experiment can replicate months of development work, what happens when that cost drops to $1.50? Or 15 cents? As data centers scale and compute costs plummet, the economics become even more brutal for software companies built on features rather than moats.

The winners in this new landscape won't be the companies with the most features — they'll be the ones with the deepest integration into workflows that can't be easily replicated. Think customer data, regulatory compliance, or network effects that compound over time.

The Bigger Questions

This experiment raises uncomfortable questions that extend far beyond software investing. If journalists can build functional apps in an hour, what happens to the 4.4 million software developers in the US? If AI can research and replicate existing tools autonomously, what's the future of incremental innovation?

The answer might lie in understanding what AI can't replicate: the deep customer insights that drive product decisions, the regulatory relationships that enable enterprise sales, or the trust that comes from years of reliable service.