In the Vibe Coding Era, Why RSS Is the Best Solution for Sharing Personal Apps

The Vibe Coding era needs RSS-like mechanisms to distribute the flood of personalized micro-apps
As Vibe Coding makes app development as simple and frequent as blogging, the centralized distribution model of traditional app stores no longer fits. Matt Webb proposes using RSS/Atom feeds to distribute these micro-apps, enabling decentralized, low-friction sharing. Simon Willison quickly put this into practice, using Claude to add an Atom feed to his tools page. While "install to where" remains an unsolved challenge, this direction reveals a macro trend of software returning to personalization and decentralization.
When Building Apps Becomes as Simple as Publishing a Blog Post
With the rise of Vibe Coding, the barrier to developing small tools and micro-applications is dropping dramatically. The concept of Vibe Coding was first introduced in February 2025 by Andrej Karpathy, former Tesla AI Director and OpenAI co-founder. It describes an entirely new approach to programming: instead of writing code line by line, developers describe their desired functionality in natural language, and large language models (such as GPT-4, Claude, etc.) generate the complete code implementation. The developer's role shifts from "code writer" to "intent expresser" and "result validator." Karpathy himself described it as "fully immersed in the vibes, embracing exponential growth, and forgetting that code even exists."
More and more developers are describing requirements in natural language and letting AI generate the code, producing usable tools in minutes. The core premise of this paradigm is the leapfrog improvement in AI code generation capabilities—especially the expansion of context windows and enhanced multi-turn conversation abilities—enabling AI to understand complex project requirements and generate structured, complete applications. This raises a new question: how should these proliferating personalized micro-applications be discovered and shared?
Matt Webb recently offered an insightful perspective on his blog: we need an RSS-like mechanism to distribute these vibe-coded applications.
How Vibe Coding Has Changed the Nature of Applications
Matt Webb wrote in his article:
When vibe-coding accelerates app development, apps become more personal, more contextual, and more frequent. Publishing a tool or micro-app no longer feels like launching a website—it feels more like publishing a blog post.
This analogy is remarkably precise. Traditional application development is a heavyweight process—requirements analysis, architecture design, development and testing, deployment and launch—each step demanding significant time and effort. But under the Vibe Coding paradigm, a person might produce several small tools every day, each solving a specific, personal problem.
This shift means that application distribution methods need to change accordingly. The centralized app distribution model represented by the Apple App Store and Google Play was established around 2008, originally designed to provide security review, payment processing, and user discovery services for complex mobile applications. This system works well for commercial applications, but presents significant structural contradictions for lightweight micro-apps: review cycles typically take days or even weeks, developers must pay annual fees and comply with complex listing requirements, and platforms take a 15%-30% revenue cut. The more fundamental issue is that these platforms' recommendation algorithms and search mechanisms are designed for a model where "a few apps serve massive numbers of users"—while the Vibe Coding era is characterized by exactly the opposite: vast numbers of highly personalized micro-apps, each potentially serving only a handful of people or even just the creator themselves.
RSS: The Underestimated Decentralized Distribution Solution
Matt Webb's proposed solution is to provide RSS/Atom feeds for these tool pages, with each entry accompanied by an "install" button.
RSS (Really Simple Syndication) and Atom are two XML-based content syndication protocols born during the golden age of blogging in the early 2000s. RSS was first introduced by Netscape in 1999 and evolved through multiple versions including 0.91, 1.0, and 2.0; Atom was standardized by the IETF in 2005 (RFC 4287) as an alternative designed to address the fragmentation of RSS specifications. Both share the same core principle: content publishers maintain a structured XML file on their server listing the latest published entries (title, summary, link, timestamp, etc.), and subscribers use RSS readers to periodically pull this file for updates. This "pull" model is inherently decentralized—no intermediary platform is needed, anyone can publish a feed, and any reader can consume it. Although RSS's mainstream usage declined after the rise of social media, it has maintained vigorous vitality in technical communities, podcast distribution, and news aggregation.
This idea may seem simple, but it captures the core of the problem:
- Decentralized distribution: Each developer maintains their own tool list and feed, independent of any platform
- Extremely low publishing friction: Publishing a new tool is as simple as publishing a blog post
- Continuous subscription: Interested people can continuously follow a developer's tool output
- Built on open standards: RSS/Atom are mature open protocols with a well-established ecosystem
Of course, Matt Webb himself raised a key question: "Install to where?" This is indeed an unresolved challenge. The industry is currently exploring multiple possible runtime environments: Progressive Web Apps (PWA) allow web applications to be "installed" to a device's home screen and run offline, making them the closest ready-made option; WebAssembly (Wasm) provides a near-native-performance browser sandbox execution environment; platforms like Val Town and Deno Deploy offer lightweight cloud-based code execution services; and Anthropic's Claude Artifacts and OpenAI's Canvas represent another approach—AI-generated code running directly within the conversation interface. Additionally, browser extensions (such as Chrome Extensions) and userscript managers (such as Tampermonkey) are potential host environments. There is no unified standard yet, which is both a challenge and an opportunity requiring further exploration from the entire ecosystem.
Simon Willison's RSS Implementation
Prominent developer Simon Willison was inspired by this article and quickly put it into action. He had Claude add an Atom feed and icon to his /elsewhere/tools/ page. This page aggregates various tools he has published on tools.simonwillison.net.
Simon Willison is one of the co-creators of the Django web framework and the author of Datasette, an open-source data exploration tool. In recent years, he has become one of the most active practitioners and evangelists in the AI-assisted programming space, particularly known for building numerous practical small tools using Claude and ChatGPT. His tools.simonwillison.net site hosts dozens of single-page web applications developed with AI assistance, covering data visualization, text processing, API debugging, and many other use cases. Each tool is typically a standalone HTML file that runs without a backend server. He also maintains the LLM command-line tool and is a significant contributor to the AI toolchain ecosystem.
A noteworthy detail: Simon himself is a deep practitioner of Vibe Coding. Using Claude to add the Atom feed was itself an instance of Vibe Coding in practice—the entire process was completed through a single GitHub PR.
A Deeper Insight: Software Is Returning to Personalization
This discussion reveals a broader trend: AI-driven development is pushing software back toward personalization and decentralization.
Over the past two decades, the dominant theme in the software industry has been platformization and centralization—a few large companies develop applications, and massive numbers of users consume them. But Vibe Coding may be reversing this trend. When everyone can easily create their own tools, what we need isn't a bigger app store, but better discovery and sharing mechanisms.
RSS solved the content distribution problem in the blogging era, and now it may play that role again in the micro-application era. This isn't a technological regression but a spiral ascent—using mature infrastructure to support entirely new use cases. Looking back at internet history, this pattern of "old protocols, new purposes" is not uncommon: HTTP evolved from a simple document transfer protocol into the foundation supporting the entire web application ecosystem; JSON went from a JavaScript subset to a universal standard for cross-language data exchange; and RSS, from a blog subscription tool to the core protocol for podcast distribution, has already proven its adaptability.
There's still a long road from RSS subscriptions to true "one-click install," but at the very least, this direction is worth serious exploration.
Key Takeaways
- Vibe coding makes app development as frequent and personal as blogging, rendering traditional app distribution models inadequate
- Matt Webb proposes using RSS/Atom feeds to distribute vibe-coded micro-apps, enabling decentralized, low-friction sharing
- Simon Willison quickly responded by using Claude to add an Atom feed to his tools page
- "Install to where" remains an unresolved key question, with PWA, WebAssembly, cloud runtime platforms, and other solutions being explored
- AI-driven development is pushing software back toward personalization and decentralization, and RSS may play an important role again in the micro-application era
Related articles
Expert OpinionsThe Lazy Person's Productivity Theory: Why Being 'Lazy' Actually Drives Peak Performance
Explore the engineering philosophy behind 'lazy people are most productive': how constructive laziness drives automation, AI tools amplify efficiency, and systems thinking eliminates wasted effort.
Expert OpinionsOutdoor Coding: You Can Touch Grass AND Build Things
When AI coding assistants free developers from their desks, outdoor coding becomes a real trend. Explore how cloud IDEs, voice coding, and AI tools enable creativity in nature.
When AI Treats Humans as Subagents: Ro…
When AI Treats Humans as Subagents: Role Reversal and Hidden Risks in Human-AI Collaboration
Exploring the paradigm shift where humans become "subagents" in AI Agent architectures. Analyzes human node design in LangChain and AutoGen, and the risks of ceding control and cognitive atrophy.