<h2>Must know</h2>
<ul>
<li><strong>Mistral Forge enables full enterprise model training</strong>: Mistral announced Forge at Nvidia GTC, a platform allowing enterprises to train custom models from scratch on proprietary data rather than relying on RAG or fine-tuning. Early partners like ASML and the European Space Agency are using it to internalize domain-specific knowledge and codebases. Mistral claims this approach reduces dependency on third-party providers and better handles non-English data. <a href="https://app.daily.dev/posts/6pejPwPmX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GNOME 50 Tokyo drops X11 and embraces Wayland only</strong>: The GNOME 50 release marks a major milestone by entirely removing X11 support to become a Wayland-only desktop environment. Key updates include official fractional scaling, VRR enabled by default, and significant NVIDIA performance fixes via Mutter. It is slated to be the default for Ubuntu 26.04 LTS and Fedora 44, as reported by GNOME Release Notes. <a href="https://app.daily.dev/posts/Xg0Exf55t" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft reshuffles AI leadership as Copilot struggles</strong>: Mustafa Suleyman is stepping back from overseeing Copilot to lead a new five-year superintelligence initiative at Microsoft. Jacob Andreou now takes over Copilot across consumer and commercial segments, reporting directly to Satya Nadella. The move follows reports that Copilot has struggled to achieve significant market impact despite aggressive integration efforts. <a href="https://app.daily.dev/posts/jTMZOA2Fr" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Edge.js runs Node.js apps inside a WebAssembly sandbox with full API compatibility</strong>: Wasmer open-sourced a runtime using WASIX that passes 99% of Node.js tests while isolating system calls. <a href="https://app.daily.dev/posts/FUA9hYuPV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Zeroboot achieves sub-millisecond VM sandboxing for AI agents</strong>: The project uses Firecracker and copy-on-write forking to launch isolated KVM sandboxes in 0.8ms. <a href="https://app.daily.dev/posts/0KZTYS81s" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Fedora Asahi Remix 43 brings Linux to M1 and M2 Mac Pro</strong>: The release includes KDE Plasma 6.6, 120Hz support for MacBooks, and microphone functionality for M2 Pro/Max. <a href="https://app.daily.dev/posts/585WUMr9y" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ConnectRPC translates GraphQL operations into typed SDKs and gRPC</strong>: WunderGraph’s new protocol layer in Cosmo Router generates Protocol Buffer definitions from named GraphQL queries. <a href="https://app.daily.dev/posts/Cgx6Z8AyK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Firefox 149 integrates a free 50GB browser-only VPN</strong>: Launching March 24, the service uses Mozilla’s own servers rather than Mullvad to protect browser-specific traffic. <a href="https://app.daily.dev/posts/nPBbJz5BW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>WordPress 7.0 standardizes AI integration with new Connectors API</strong>: The framework provides a unified way to manage API keys and provider discovery for Anthropic, Google, and OpenAI. <a href="https://app.daily.dev/posts/6us9IvcEV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GRUB development migrates to FreeDesktop GitLab for modern workflows</strong>: The project is moving away from GNU Savannah to adopt a merge-request-based contribution model. <a href="https://app.daily.dev/posts/EiDnSja30" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The prompt injection arms race is now a $2 per million token tax</strong>: Arcjet launched a ‘protection’ feature that uses a purpose-trained LLM to check your prompts for malicious intent before they hit your actual LLM. It’s a classic case of solving the problems created by AI by adding more AI, introducing 100-200ms of latency and a recurring bill just to stop users from asking your chatbot to act like a pirate. <a href="https://app.daily.dev/posts/B5yvripTT" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Sauce Labs wants to fix the AI bottleneck with more AI</strong>: After AI code generation flooded the world with untested garbage, Sauce Labs is pitching ‘AI for Test Authoring’ to generate test suites from Jira specs. It’s a beautiful cycle of inefficiency where we use one LLM to write code we don’t understand and another to write tests that might not actually prove it works. <a href="https://app.daily.dev/posts/DdGc4GnFc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Alibaba Cloud blames the environment for its 34 percent price hike</strong>: Alibaba is jacking up prices for GPU instances by a third, citing ‘global AI demand’ and ‘hardware costs’ while simultaneously bragging about their superior resource efficiency. It’s a masterclass in opportunistic pricing disguised as a supply chain crisis, effectively rationing compute to whoever has the deepest pockets. <a href="https://app.daily.dev/posts/8JNWgjxHp" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
