The way we use software is on the verge of a fundamental shift. For decades, users have been forced to adapt to the features and designs of commercially built applications. If a program didn't do exactly what you needed, your only option was to learn to code or simply make do. That era is ending.
With the rapid advancement of generative AI models, particularly tools like Anthropic's Claude Code, OpenAI's Codex, GitHub Copilot, and others, the ability to create functional software is now accessible to anyone with an internet connection and a few dollars a month. This phenomenon, dubbed 'vibe coding' by AI researcher Andrej Karpathy, allows people to describe an app idea in plain language and have an AI generate the code. The result is a surge in what many call 'personal software' — applications built for an audience of exactly one.
The core promise of personal software is liberation from the tyranny of one-size-fits-all products. Instead of cobbling together workarounds using Apple Shortcuts or IFTTT, users can now build bespoke tools that fit their exact workflow. For example, a family can create a custom meal planner with a built-in grocery assigner for their next trip. A migraine sufferer can build a dedicated tracker to log symptoms and triggers. Even things like tracking which stair your mail carrier leaves a package — a problem for someone with a long staircase — can now have its own dedicated app. These tools rarely generate revenue or attract venture capital; they are purely functional, serving the creator's specific needs.
The trend has already caused measurable changes in the software ecosystem. According to reports, the number of new apps in Apple's App Store grew by 30% in 2025, reversing nearly a decade of decline. GitHub also saw its fastest year of growth, with 80% of new users adopting the Copilot coding agent within their first week. Much of this growth is driven by 'barefoot developers' — people who learn just enough coding to solve their own problems, often with heavy AI assistance.
Yet the movement has clear limitations. AI-generated apps often suffer from poor design, security vulnerabilities, and lack of support. Many users find themselves trapped in 'doom loops,' endlessly telling the AI what they don't like without a clear vision of what they want. The AI's default aesthetic often leans toward purple gradients and hamburger menus, frustrating those with refined taste. Furthermore, while building a prototype is fast, making the app robust, syncing across devices, and handling edge cases takes significant effort and technical knowledge.
Nevertheless, the philosophy of personal software is resonating deeply. Robin Sloan, an author and technologist, famously wrote about building a home-cooked messaging app for his family in 2020, long before the current AI boom. He updated the post in late 2025, noting that he had changed nothing about the app in five years, and it remained perfect. Today, Sloan uses AI to create small, hacky scripts that automate tasks like generating shipping labels for his olive oil company. 'It's always weird little things,' he says. 'If I ever get hit by a bus, it’s going to be a problem for my olive oil company, because only Robin knows how to run the software.' But while he's around, it works great.
The security and reliability concerns are real. Most personal apps are not tested for vulnerabilities, do not offer data backups, and have no customer support. For mission-critical enterprise use, professionally built software remains essential. However, for the small, annoying problems that commercial software solves poorly, personal AI-generated apps are becoming the go-to solution.
This new paradigm also challenges traditional developers. Companies like Notion are building platforms that allow users to customize their workspace using AI to 'write macros' rather than full software. The CEO of Craft, a note-taking app, emphasizes that while customization is powerful, providing a coherent out-of-the-box experience remains crucial. Developers must now think about how to make their products both powerful and safe to customize, preventing users from accidentally breaking their software.
In the end, the most important skill for a personal software creator is taste — the ability to articulate what you want, not just what you don't. As one designer put it, 'The confidence I have in my taste, and my ability to express what I feel' is what makes the difference between a successful app and a frustrating failure. For those willing to learn a few prompts and experiment, the reward is software that works exactly the way they want. No coding bootcamp required, just a subscription to Claude Code and a clear idea of what you need.
The revolution is not about replacing professional developers. It's about empowering individuals to build the tools they wish existed. From fantasy baseball ranking scripts to renewable energy mods for 1990s games, the creativity unleashed by AI is producing an explosion of niche, personal software. The total addressable market for most of these apps is one person, and the revenue potential is zero dollars. But in an industry obsessed with scale and growth, that might be the most liberating change of all.
Source: The Verge News