Andrej Karpathy coined the term in early 2025. By the end of that year, Collins Dictionary named it Word of the Year. By February 2026, 92% of US developers use AI coding tools daily, and GitHub reports 46% of all new code is now AI-generated.
Vibe coding won. It genuinely, decisively won the adoption argument. Nobody serious is debating whether you should use these tools anymore.
The debate that actually matters is the one the community keeps avoiding: winning the build sprint and losing the business is still losing.
Shipping became the dopamine hit.
The thing vibe coding did is real and worth acknowledging clearly. It collapsed the distance between an idea and a working prototype from months to days, sometimes hours.
Pieter Levels built a browser-based 3D flight simulator in three hours using Cursor AI, ThreeJS, and Claude, despite having zero game development experience.
Within ten days it was generating over $87,000 a month from in-game advertising. That story is not an exaggeration and it is not reproducible at scale, which is exactly the problem with how it gets told.
The Levels story spread everywhere because it confirmed what every developer who hates sales wanted to believe: that if you build something interesting enough, the revenue finds you.
The reality is that Levels had 400,000 Twitter followers before he wrote a single line of that flight simulator. He didn’t find customers. He already had an audience and he gave them something to pay for.
Most people celebrating the vibe coding revolution don’t have 400,000 followers. They have a working app and a Product Hunt page and a vague plan to “do marketing later.”
There is no later. There was only ever the customer.
The finish line changed
Every tool in the vibe coding stack, Cursor, Claude, Bolt, Lovable, Supabase, Vercel, is optimized for one outcome: getting something live. That is what the tools reward.
That is what the build-in-public culture on X rewards. Shipping screenshots, MRR day-one announcements, “I built this in a weekend” posts. The entire feedback loop from idea to deploy is now fast, frictionless, and publicly celebrated.
The feedback loop from idea to paying customer still runs on completely different physics.
Customers don’t care that you shipped in 48 hours. They care whether the thing solves a problem they were already trying to solve before you built it. User interviews are not a vibe.
Cold outreach is not a vibe. Sitting on a call with someone who has no interest in buying and figuring out why, then using that to reshape your positioning, is deeply anti-vibe.
None of those activities show up in a build thread. None of them feel like progress on the day you do them.
So people skip them. They build another app instead, because building feels like forward motion and selling feels like failure pending.
Hacker News discussions surfaced businesses now offering “vibe coding cleanup as a service” specifically to fix disasters left behind, with consultants noting the cleanup cost often exceeds the cost of proper development from scratch.
The same pattern is playing out at the business level. The cleanup cost of launching nine apps without a single customer conversation is not technical debt. It is months of misallocated attention that compound into a habit of shipping without validating.
The security problem is a symptom of a deeper one
The Moltbook breach is the case study nobody in the vibe coding community wants to dwell on. The AI-agent social network launched January 28, 2026, with the founder proudly stating he did not write one line of code.
Three days later, security researchers found 1.5 million API authentication tokens, 35,000 email addresses, and private messages containing plaintext credentials exposed. The root cause was a Supabase API key sitting in client-side JavaScript with Row Level Security not enabled. bloomingbit
The community responded by writing checklists. Security audits before launch. Environment variable hygiene. Backend paywalls instead of frontend logic.
All useful, all necessary, and all completely missing the real lesson.
The real lesson is that building something you don’t understand at a layer of depth puts you in a position where you cannot tell the difference between a product that works and a product that appears to work.
Those are not the same thing. One generates revenue and compounds over time. The other generates launch dopamine and then a shutdown post six months later.
An indie developer built an entire SaaS product with Cursor, zero hand-written code, celebrated on social media when users signed up, and then watched it collapse when random things started happening, API keys got maxed out, people were bypassing the subscription, and the database was filling with garbage.
He couldn’t debug it because he didn’t write it. The product was shut down permanently.
Vibe coding didn’t cause that failure. The belief that launching was the hard part did.
The winners are using it differently than the losers
There is a real pattern among the vibe coders who are generating sustainable revenue, and it has nothing to do with which AI tool they’re using or how fast they can ship.
They validated before they built. They found a specific person with a specific problem, confirmed that person was already trying to solve it unsuccessfully, and then used vibe coding to build the solution faster than a traditional developer could.
The speed was an advantage. The customer was the strategy.
The losers in this story, and there are genuinely many of them posting memes about nine apps and zero revenue, reversed the order.
They built first because building is the fun part, then went looking for a problem their solution might fit, then got confused when the market didn’t respond to something it was never asked for.
The winners in 2026 are founders who direct AI agents with clear intent, not those who generate the most code. That sentence should be printed on every Cursor onboarding screen.
What the culture actually needs to celebrate
The build-in-public movement documents shipping obsessively and documents customer development almost not at all. Nobody posts about the twelve user interviews that revealed their core assumption was wrong. Nobody threads their cold email sequence and the 94% non-response rate they pushed through before finding a pattern.
Those stories are not shareable in the same way. They don’t generate the same engagement. But they are the actual work, and the absence of those stories from the cultural feed is creating a deeply distorted picture of what building a business requires.
Vibe coding gave an entire generation of developers genuine superpowers on the production side of building software. The market does not grade on production speed. It grades on whether you found a real customer who pays you real money to solve a real problem, and whether you can keep doing that at a scale that sustains itself.
The finish line was always the customer. The tools just made it easier to run very fast in the wrong direction.
Comments
Loading comments…