<h2>Must know</h2>
<ul>
<li><strong>GNOME 50 Tokyo drops X11 and embraces Wayland only</strong>: Released on March 18, 2026, GNOME 50 marks a definitive end to X11 support, transitioning to a pure Wayland environment. This milestone includes official fractional scaling, Variable Refresh Rate support, and significant NVIDIA performance fixes via enhanced frame scheduling. According to GNOME Release Notes, the update also introduces GPU-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>Mistral Forge enables full enterprise model training</strong>: Mistral announced Forge at Nvidia GTC, a platform allowing organizations to train custom AI models from scratch on proprietary data rather than relying on RAG or fine-tuning. The platform targets high-stakes sectors like government and finance, offering support for both dense and mixture-of-experts architectures. Mistral claims Forge is agent-first, enabling autonomous systems to manage training jobs using natural language. <a href="https://app.daily.dev/posts/6pejPwPmX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Erlang OTP 29.0 RC2 introduces quantum resistant SSH</strong>: The second release candidate for OTP 29.0 is now available, switching the default SSH key exchange to a quantum-resistant algorithm. Key technical additions include the io_ansi module for terminal sequences and native records as an experimental feature. The update also brings JIT improvements for binary matching and new compiler warnings for deprecated operators. <a href="https://app.daily.dev/posts/43R5cEvzp" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Fedora Asahi Remix 43 brings Linux to M1 and M2 Macs</strong>: The release features KDE Plasma 6.6, 120Hz support for MacBook Pros, and microphone functionality for M2 Pro/Max models. <a href="https://app.daily.dev/posts/585WUMr9y" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Zeroboot achieves sub millisecond VM sandboxing for agents</strong>: Using Firecracker and copy-on-write forking, this prototype forks pre-snapshotted VMs in 0.8ms with minimal memory overhead. <a href="https://app.daily.dev/posts/0KZTYS81s" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AlmaLinux 10 begins supporting RISC-V hardware</strong>: The testing branch now produces RV64GC builds with upcoming support for VisionFive 2 and HiFive Premier P550 boards. <a href="https://app.daily.dev/posts/3a3YENAAh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GRUB development migrates to FreeDesktop GitLab</strong>: The project is moving away from GNU Savannah to adopt a modern merge request workflow for future contributions. <a href="https://app.daily.dev/posts/EiDnSja30" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Redpanda opensources Go AI SDK with MCP support</strong>: The SDK provides provider portability across major LLMs and includes a simulated framework for deterministic testing. <a href="https://app.daily.dev/posts/8GYlam0VJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ConnectRPC translates GraphQL to typed SDKs</strong>: A new protocol layer in Cosmo Router compiles GraphQL operations into Protobuf definitions for gRPC and OpenAPI generation. <a href="https://app.daily.dev/posts/Cgx6Z8AyK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AMD MLIR AIE v1.3 adds native C++ compiler</strong>: The new aiecc compiler replaces Python tooling, offering C++17 support and better instruction generation for Ryzen AI NPUs. <a href="https://app.daily.dev/posts/gLtp8gAqE" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The AI productivity tax is real and it is 19 percent</strong>: While marketing departments scream about 70% time savings, actual studies from SmartBear show developers facing a 19% overall slowdown due to context-switching and a 50% spike in debugging time. We’ve traded writing code for babysitting a hallucinating intern who occasionally deletes production databases. It turns out ‘vibecoding’ doesn’t scale when you actually have to ship working software. <a href="https://app.daily.dev/feed-by-ids?id=Lb19kFYlj&amp;id=gDgGS8RRB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Copilot is so good they had to stop deploying it</strong>: Microsoft has ‘paused’ the automatic rollout of the 365 Copilot app after IT admins and commercial customers collectively lost their minds over the forced opt-out strategy. Nothing says ‘world-class product’ like a deployment so aggressive it causes a global backlash from the people paid to manage it. It’s a classic case of a solution in search of a problem, now indefinitely on ice. <a href="https://app.daily.dev/feed-by-ids?id=w3Je7Auua&amp;id=jTMZOA2Fr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Nvidia networking is the new Cisco but with more hype</strong>: Nvidia’s networking division is now out-earning Cisco’s entire business, fueled by the desperate need to connect thousands of GPUs that mostly generate slop. While they announce ‘Rubin’ platforms and ‘Spectrum-X’ switches, the reality is a proprietary lock-in cycle that makes the old Cisco days look like an open-source utopia. We’re just swapping one networking overlord for another with a better stock ticker. <a href="https://app.daily.dev/feed-by-ids?id=6uX3MWeQL&amp;id=LtvttiElC" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
