We need RSS for sharing abundant vibe-coded apps

The Catalyst: Why Discuss an 'RSS for Apps' Now? Simon Willison, quoting Matt Webb, raises a seemingly simple yet profoundly forward-looking question: How do we share and manage apps when AI makes creating them as easy as writing a blog post? The context for this question is the rise of 'vibe-coding.' With large language models, developers (and even non-developers) can quickly generate a usable tool or micro-app based on a vague idea or a 'feeling.' This represents a fundamental shift in the speed and volume of app creation. In the past, publishing an app was a major undertaking involving design, development, testing, and deployment. Now, it can be the product of an afternoon's 'vibe.' When apps transform from 'carefully crafted products' to 'casually published digital artifacts,' the existing app store model becomes obsolete and cumbersome. Deconstruction: What is the Core Idea? The core insight here is that the nature and unit of distribution for apps are changing. 1. Personalized and Situated Apps: These AI-assisted, rapidly generated apps are often deeply tied to the creator's specific needs, workflows, or momentary inspiration. They might be a small tool for handling a particular data format, a script automating a mundane task, or a prototype validating an idea. They resemble personal notes or blog posts more than traditional software, carrying a strong personal context. 2. A Distribution Model Analogous to Blogging: Matt Webb brilliantly analogizes: "Shipping a tool or micro-app is less like launching a website and more like posting on a blog." The heart of blog distribution is RSS/Atom—a decentralized, open protocol based on user-initiated subscriptions. Readers aggregate information through feed readers, not via a centralized platform. Willison's action validates this idea: he directly had an AI (Claude) add an Atom feed to his tools page. This itself is a perfect example of 'vibe-coding,' demonstrating through action how simple it can be to add a subscription feature to an app. Trend Insight: What Larger Trend Does This Reveal? This reveals a deep evolution in software form and distribution paradigms in the AI era. * From 'Software as a Service' to 'Software as Content': When the marginal cost of creating software approaches zero, software itself may become more akin to a form of 'content' or 'information.' Its lifecycle might be short, updates very frequent, and its value lies in solving an immediate problem rather than long-term maintenance. This shares similarities with blog posts or social media updates. * A Renewed Need for Decentralized Distribution: Centralized app stores are suitable for distributing serious, commercial software products. However, for the flood of personal, experimental 'vibe-coded' apps, a lightweight, open, decentralized subscription protocol (like RSS) might be a more natural fit. It allows creators to maintain control over distribution channels and lets users subscribe on demand, avoiding information overload. * The Rise of 'Meta-Tools': Tools for managing these apps will become a new necessity. Just as feed readers are used to consume blogs, we might need 'app readers' or 'personal app dashboards' in the future to discover, try, update, and uninstall the stream of micro-apps being produced. Practical Value: How Can Readers Think, Use, or Judge? For IT/internet professionals, this trend is worth considering from several angles: 1. As a Creator: If you start using AI to rapidly build internal tools or prototypes, consider adding a simple RSS/Atom feed to your 'portfolio' page. This is not just about sharing; it's a way to build a personal tech brand and attract like-minded individuals. Willison's example shows that the technical barrier to implementing this is now extremely low. 2. As a Team Manager: Within enterprises, such 'vibe-coded' apps may already be emerging quietly. Thinking about how to establish an internal 'app catalog' or subscription mechanism so that valuable micro-tools can be discovered and reused by the team, rather than being scattered across individual local environments, could be a new lever for improving organizational efficiency. 3. As an Observer: Watch for innovations in the surrounding ecosystem for distributing, managing, and securing (how do you trust a casually generated app?) AI-generated applications. This could be the next blue ocean for toolchain startups. Counterintuitive/Overlooked Angle: What Might Most People Miss? An easily overlooked point is the blurring of the concept of 'installation.' Matt Webb asks in his quote: "(But install to where?)" This question is profoundly insightful. Traditional app installation implies a persistent binary file residing in a specific location on a device. A 'vibe-coded' app, however, might be: a script that can be regenerated at any time, a temporary service existing only in the cloud, or a code snippet that requires a specific context to run. Its 'form of existence' is fluid. Therefore, future 'subscriptions' might not just be about obtaining an installer package, but about acquiring a 'capability description' or 'generation instruction,' allowing users to 'instantiate' the app whenever needed. This points towards a lighter, more dynamic model of software consumption.