<h2>Must know</h2>
<ul>
<li><strong>GNOME 50 drops X11 and goes all in on Wayland</strong>: The ‘Tokyo’ release marks a definitive end to X11 support, transitioning GNOME to a Wayland-only environment. This version graduates fractional scaling and variable refresh rate (VRR) to stable defaults while significantly improving NVIDIA GPU frame scheduling. According to release notes, the update also brings hardware-accelerated remote desktop sessions via Vulkan and VA-API. <a href="https://app.daily.dev/posts/Xg0Exf55t" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Wasmer Edge.js brings Node.js apps to WebAssembly</strong>: Wasmer has launched Edge.js, a new JavaScript runtime that allows unmodified Node.js workloads to run inside a WebAssembly sandbox using WASIX. While it preserves full Node.js v24 API compatibility, the sandboxing currently introduces a 5–30% performance overhead compared to native execution. The project aims to provide a secure, isolated environment for AI and edge computing tasks without the code changes required by Deno or Cloudflare Workers. <a href="https://app.daily.dev/feed-by-ids?id=qHzldup0J&amp;id=FUA9hYuPV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Systemd 260 introduces AI coding guidelines and controversy</strong>: The latest Systemd release officially drops support for legacy System V init scripts and introduces an <a href="http://AGENTS.md" target="_blank" rel="noopener nofollow">AGENTS.md</a> file specifically to guide AI coding agents. The project now mandates human review and explicit disclosure for AI-generated code in commit messages following a controversial sd-bus contribution assisted by Claude. This move has landed Systemd on the ‘OpenSlopware’ list, sparking debate over AI’s role in critical system plumbing. <a href="https://app.daily.dev/posts/ynJvFeH96" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Laravel 13 stabilizes AI SDK and adds vector search</strong>: The new release requires PHP 8.3 and introduces native semantic vector search for PostgreSQL alongside a stabilized AI SDK. <a href="https://app.daily.dev/posts/nHKWQwLUI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Zeroboot achieves sub-millisecond VM forking for AI agents</strong>: Using Firecracker and KVM copy-on-write forking, this project can spin up isolated sandboxes in 0.8ms with only 265KB memory overhead. <a href="https://app.daily.dev/posts/0KZTYS81s" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>RubyLLM 1.14 scaffolds full AI chat UIs in minutes</strong>: New Rails generators automate the creation of Tailwind-based chat interfaces, agent tools, and schemas following standard conventions. <a href="https://app.daily.dev/posts/4PqMrUSSr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Fedora Asahi Remix 43 brings Linux to M1 and M2 Macs</strong>: The update adds support for Mac Pro models, 120Hz refresh rates on MacBook Pros, and microphone functionality for M2 Pro/Max chips. <a href="https://app.daily.dev/posts/585WUMr9y" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ConnectRPC enables typed SDK generation from GraphQL</strong>: The Cosmo Router now supports transcoding GraphQL operations into gRPC and Connect protocols to generate native Go, TS, and Swift clients. <a href="https://app.daily.dev/posts/Cgx6Z8AyK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Java 26 adds HTTP/3 support and AOT object caching</strong>: The non-LTS release focuses on performance with JEP 517 (QUIC) and JEP 516, which improves object caching for ZGC. <a href="https://app.daily.dev/feed-by-ids?id=26rrhGWJk&amp;id=QPwbRCIfz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Mystral Native.js runs WebGPU apps without a browser</strong>: This early-alpha runtime uses V8 and SDL3 to execute JavaScript games natively on desktop with direct WebGPU and Web Audio access. <a href="https://app.daily.dev/posts/YUabTxSP2" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Meta abandons the Metaverse for the mobile screen</strong>: After burning billions in Reality Labs, Meta is unceremoniously killing Horizon Worlds on Quest VR headsets to focus on iOS and Android apps. It turns out the ‘future of work and play’ is just another mobile game you’ll ignore on your phone. The pivot to AI and smart glasses is the new shiny object to distract investors from the smoking crater of their VR ambitions. <a href="https://app.daily.dev/posts/etikjr6BR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft reshuffles AI leadership as Copilot stalls</strong>: Mustafa Suleyman is being moved to a ‘superintelligence’ role—the corporate equivalent of being sent to a farm upstate—while Jacob Andreou tries to fix Copilot’s lack of market impact. Despite shoving AI into every context menu and taskbar, Microsoft is finding that aggressive integration doesn’t equal actual user utility. Calling it a ‘superintelligence initiative’ is a classic way to hide that your current product is struggling to justify its compute costs. <a href="https://app.daily.dev/posts/jTMZOA2Fr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google DeepMind wants to benchmark ‘meatbags’</strong>: DeepMind is launching a $200,000 Kaggle hackathon to create a ‘cognitive taxonomy’ for AGI, using sensationalist language to frame it as humans vs. machines. In reality, it’s just a desperate search for new benchmarks because current LLMs have already memorized all the existing ones. If they can’t pass the bar, they’ll just keep moving the bar until the marketing department is satisfied. <a href="https://app.daily.dev/posts/NhVe5iupm" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
