<h2>Must know</h2>
<ul>
<li><strong>Tailwind CSS revenue collapse due to AI documentation use</strong>: Tailwind CSS reported a 75% staff layoff and an 80% revenue drop, despite increased user adoption. The company attributes this to AI models directly answering user queries, bypassing their documentation site which is crucial for converting users to paid products. This highlights a significant challenge for open-source projects relying on documentation-based monetization, as AI consumes content without contributing to the project’s sustainability, as reported by Tailwind CSS itself and various industry analyses. <a href="https://app.daily.dev/feed-by-ids?id=RuMhQtj6W&amp;id=f3fYMXuKZ&amp;id=WagqERKHJ&amp;id=eugOrAHSX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>VS Code introduces agent skills for AI agents</strong>: Visual Studio Code (VS Code) has rolled out an experimental feature called Agent Skills, allowing AI agents like GitHub Copilot to dynamically load specialized instructions and resources for specific developer tasks. This aims to enhance automation and context-aware workflows, enabling developers to teach Copilot new abilities. The feature uses an open standard, ensuring portability across compatible AI agents, according to official VS Code communications. <a href="https://app.daily.dev/posts/xn90KJxf1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Claude Code 2.1 boosts developer workflows</strong>: Anthropic has released Claude Code 2.1, introducing significant enhancements for developers. Key updates include forked and asynchronous sub-agents for parallel task execution, automatic skill hot reloading, and session teleportation between terminal and desktop environments. It also adds Chrome browser integration for web testing and persistent memory across sessions, aiming to optimize developer productivity and efficiency. <a href="https://app.daily.dev/feed-by-ids?id=WChBmOscl&amp;id=nDBQDaZMg" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>New open-source React scheduler component released</strong>: CalendarKit, a new open-source React calendar and scheduler component library, is now available in Basic and Pro versions, offering features like drag-and-drop, dark mode, TypeScript support, and recurring event handling for various application integrations. <a href="https://app.daily.dev/feed-by-ids?id=MBDqbsDnk&amp;id=RcOUV5N3R&amp;id=mf2DGEww0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel Boost updates for AI agent compatibility</strong>: Laravel Boost v1.8.9 now supports the Model Context Protocol (MCP), dropping legacy protocol compatibility to align with new AI agents like Claude Code, requiring users to upgrade to maintain compatibility. <a href="https://app.daily.dev/posts/8827svTqM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WebGPU updates in Chrome 144</strong>: Chrome 144 introduces new WGSL language extensions for WebGPU, including <code>subgroup_id</code> and <code>uniform_buffer_standard_layout</code>, alongside performance improvements and the first alpha of Kotlin bindings for Android GPU. <a href="https://app.daily.dev/posts/8boXOlRd2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Kubernetes database management platform OpenEverest</strong>: Percona has donated its database management platform, now rebranded as OpenEverest, to the CNCF, providing a unified interface for managing PostgreSQL, MySQL, and MongoDB on Kubernetes with a web interface and RESTful API. <a href="https://app.daily.dev/posts/MhqrMxmeE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel ecosystem health directory launched</strong>: <a href="http://LaraPlugins.io" target="_blank" rel="noopener nofollow">LaraPlugins.io</a> is a new directory that helps developers find and evaluate Laravel packages using an automated health scoring system based on Composer and GitHub metrics, ensuring well-maintained dependencies. <a href="https://app.daily.dev/posts/d8XfhgQdf" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Orion browser for Linux in alpha testing</strong>: The Orion browser for Linux is currently in alpha, with core features like tab management, session persistence, bookmarks, and password management already implemented, and WebKit extension support planned. <a href="https://app.daily.dev/posts/fMKCRJ6El" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Postman acquires Fern for API documentation and SDK generation</strong>: Postman has acquired Fern, a company specializing in API documentation and SDK generation tools, to enhance its developer experience offerings and expand its capabilities in API collaboration. <a href="https://app.daily.dev/posts/YZCRLW376" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI code quality: more code, more problems?</strong>: SonarSource’s latest leaderboard shows AI models like GPT-5.2 High generating three times more Java code than others, but with a higher issue density, including SQL injection risks. Apparently, the latest LLM versions don’t always improve code quality, which is a real shocker for anyone who’s been paying attention. So, while 72% of developers use AI daily, 96% still don’t trust its output, and 48% don’t even bother to verify it before committing. What could possibly go wrong? <a href="https://app.daily.dev/feed-by-ids?id=E02dc9OJZ&amp;id=q2sjuwRsf" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents: the new frontier of cloud security risks</strong>: Security experts are now warning that AI agents, with their ‘lack of human judgment,’ are exposing legacy security gaps by attempting every possible accessible action. The Model Context Protocol (MCP), while enabling AI agents to access tools efficiently, apparently ‘increases security risk by widening context windows and uncontrolled visibility.’ So, we’re building smarter agents that are just better at finding all the ways we messed up our security. Progress! <a href="https://app.daily.dev/posts/lqqfnNpcw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI in software development: faster to fail?</strong>: A software consultancy found that while AI tools speed up early ideation, they also create ‘anchoring effects’ that limit innovative thinking and ‘review bottlenecks’ because AI generates changes faster than humans can understand them. This leads to ‘knowledge erosion’ among developers, making software harder to maintain. So, we’re trading deep understanding for rapid iteration, which sounds like a fantastic recipe for hidden, partial failures that go unnoticed for weeks, as noted by various technical practitioners. <a href="https://app.daily.dev/feed-by-ids?id=DRTczTJyG&amp;id=K8Q8uNU3E&amp;id=Kf8FXuyo" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
