Vibe coding: When everyone can build software – and what it means for you
It’s been exactly one year since the term emerged. Yet the entire tech industry is talking about it, and the platforms behind it are growing faster than anything we’ve seen in a long time. Vibe coding isn’t just another tech phenomenon drifting past in your LinkedIn feed. It’s a real shift in who can build software – and it challenges something fundamental about the way we work with digital solutions.
From code to conversation
On 6 February 2025, Andrej Karpathy – co-founder of OpenAI and former head of AI at Tesla – posted a tweet that quickly spread through the tech world. He described how he now lets AI write all his code without even reading it. When he gets an error message, he pastes it directly into the AI without any further explanation. And it works – most of the time.
That’s the essence of vibe coding: you describe in plain language what you want built, and an AI generates the code for you. No syntax, no programming logic, no technical background required. You still set the direction – but you leave the actual coding to the machine.
Karpathy had actually pointed in this direction as early as 2023, when he said that “the hottest new programming language is English.” Back then it sounded almost like a joke. Today it’s a business model.
Why is it exploding now?
Speed is the most obvious driver. You can go from idea to working prototype in minutes – not weeks. An internal dashboard, an analytics tool or a campaign landing page is suddenly something you can have ready the same day you get the idea.
Accessibility is the other major driver. A marketing professional, a project manager or an entrepreneur can now create the solutions they need – without waiting for the IT department to have capacity. The CEO behind Lovable puts it plainly: “Only one percent of the world’s population can code. The other 99 percent are my target audience.”
And then there’s the experimentation value. Because building prototypes is fast and cheap, you can test ideas that would otherwise never have been prioritised – and show a proof of concept to clients or management before investing in real development.
It’s not all summmer and sunshine…
Security is the most serious problem. In May 2025, a study found that 170 out of 1,645 apps built with Lovable had security vulnerabilities that allowed personal data to be accessed by anyone. AI generates code it doesn’t necessarily understand from a security perspective – and if you push it to production without a technical review, you could end up with a solution that’s vulnerable to even basic attacks.
Code quality is another issue. An analysis of nearly 500 open source projects from December 2025 found that AI-written code contained twice as many critical bugs as human-written code – and security vulnerabilities were three times more common. On top of that comes technical debt: code written quickly becomes increasingly expensive to maintain over time. The larger a project gets, the harder it is to manage with natural language alone – context grows, patterns become inconsistent, and you end up with a codebase no one really understands.
The bottom line is clear: vibe coding is great for prototypes and experiments. But never push it to production – especially if it handles personal data – without having an experienced developer review it first.
The platforms: what are the options?
The market is already surprisingly mature and can be divided into two categories.
For those without a technical background, Lovable and Bolt.new are the most obvious starting points. Lovable has one of the wildest growth stories in European tech: from one million to one hundred million dollars in revenue in eight months, and a valuation of 6.6 billion dollars following a funding round in late 2025. You describe what you want to build, and Lovable delivers – frontend, backend and database. Bolt.new runs directly in the browser and is the fastest way from idea to something you can show others. Google AI Studio is a free alternative within Google’s ecosystem, where you can deploy directly to the cloud with a single click.
For those with some technical experience, Cursor is the most widely used option – a code editor with AI built directly into the workflow, used by seven million developers globally. The typical approach is to start in Lovable or Bolt and then move the project over to Cursor when it needs to become something serious. Claude Code and OpenAI’s Codex are more advanced engineering agents, primarily relevant for dedicated development teams.
What does it mean for you as a marketer?
As a marketing professional, you can use vibe coding to build dashboards and reporting tools tailored to exactly the data you need – without waiting for IT. You can test landing pages for campaigns before investing in proper development, and build internal automation tools that save your team from repetitive manual tasks.
And perhaps least obvious but genuinely valuable: vibe coding makes it easier to communicate with developers. A working prototype is far more precise than a Word document or a wireframe.
But remember that GDPR applies regardless of whether the code was written by a human or an AI – if your solution handles personal data, a security review is required before it goes live..
Vibe coding is already becoming outdated
The interesting thing is that the man who coined the term has already moved on. In February 2026, Karpathy introduced the concept of “agentic engineering” and argued that the new standard is for professionals to not write code at all for the majority of their time – instead orchestrating AI agents and overseeing quality as engineers rather than craftspeople.
It’s not about prompting blindly and accepting all output. It’s about directing AI with professional judgement and ensuring the result meets the quality you’d expect.
Whatever you call it, the direction is clear: the barrier to building software is falling. That opens doors for new ways of working and new ways of solving concrete challenges – including in marketing. The term is only one year old. Just look at what’s happened since.
