<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. This ‘agent-first’ platform targets high-stakes sectors like finance and government, offering forward-deployed engineers to assist with data pipelines. Early partners include ASML and the European Space Agency, signaling a shift toward deep model ownership for the enterprise. <a href="https://app.daily.dev/posts/6pejPwPmX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents double credential leak rates on GitHub</strong>: GitGuardian’s 2026 State of Secrets Sprawl report reveals that AI coding tools have doubled the rate of credential leaks, with Claude Code-assisted commits leaking secrets at twice the baseline rate. Despite the surge, 64% of valid secrets leaked years ago remain unrevoked, highlighting a systemic governance failure as agents gain autonomous API access. Security researchers at Oasis also disclosed ‘Claudy Day,’ a vulnerability chain allowing silent data exfiltration via the Anthropic Files API. <a href="https://app.daily.dev/feed-by-ids?id=3ezJkKAUp&amp;id=4tcK5PNQ1&amp;id=54LR4cpls" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GNOME 50 Tokyo drops X11 for a Wayland only future</strong>: The GNOME 50 release officially removes X11 support, completing the transition to a Wayland-only desktop environment. Key technical updates include official fractional scaling, VRR enabled by default, and GPU hardware acceleration for remote desktop sessions via Vulkan. This version will be the default for Ubuntu 26.04 LTS, bringing significant NVIDIA performance fixes and improved discrete GPU detection to the mainstream Linux desktop. <a href="https://app.daily.dev/posts/Xg0Exf55t" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>ConnectRPC translates GraphQL to type safe SDKs</strong>: Cosmo Router now compiles GraphQL operations into Protobuf definitions, serving gRPC and generating typed Go or TypeScript clients without schema changes. <a href="https://app.daily.dev/posts/Cgx6Z8AyK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Zeroboot achieves sub millisecond VM sandboxing</strong>: Using Firecracker and copy-on-write forking, this prototype forks pre-snapshotted VMs in 0.8ms for isolated AI agent execution. <a href="https://app.daily.dev/posts/0KZTYS81s" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Edge.js runs Node apps in WebAssembly sandboxes</strong>: Wasmer’s new runtime preserves full Node.js v24 API compatibility while isolating syscalls via WASIX, passing 99% of the official test suite. <a href="https://app.daily.dev/posts/FUA9hYuPV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Bun 1.3.11 adds native cron and ANSI aware slicing</strong>: The update introduces Bun.cron for OS-level scheduling and fixes critical http2 stalling and Windows-specific fs bugs. <a href="https://app.daily.dev/posts/DuG54Q8tB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft reshuffles AI leadership as Copilot struggles</strong>: Mustafa Suleyman is pivoting to a ‘superintelligence’ initiative while Jacob Andreou takes over Copilot following weak market impact. <a href="https://app.daily.dev/posts/jTMZOA2Fr" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Fedora Asahi Remix 43 brings Linux to M2 Pro Max</strong>: The new release supports 120Hz refresh rates on MacBook Pros and introduces microphone support for M2 series hardware. <a href="https://app.daily.dev/posts/585WUMr9y" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Firefox 149 bundles a free 50GB browser VPN</strong>: Launching March 24, the built-in VPN uses Mozilla’s own infrastructure to protect browser-only traffic in the US and Europe. <a href="https://app.daily.dev/posts/nPBbJz5BW" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The 200k hallucination tax</strong>: Waterline Development managed to burn $200,000 and four months of research because they trusted ‘confidently incorrect’ answers from Grok and ChatGPT for materials science. Their solution is Rozum, yet another ‘orchestration layer’ that runs an ensemble of models to check if the first one is lying. It’s a classic case of paying three times the price to fix a problem that wouldn’t exist if we just used a library and a brain. <a href="https://app.daily.dev/posts/yxs1ZEOjQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Nvidia networking is the new Cisco</strong>: Nvidia’s networking division, a side quest from the Mellanox acquisition, just out-earned Cisco’s entire networking business in a single quarter with $11 billion in revenue. While Cisco tries to stay relevant by sticking Blackwell GPUs into their rack servers, Nvidia is busy owning the entire data center fabric. It’s not a partnership; it’s a controlled demolition of the legacy networking market disguised as ‘collaboration.’ <a href="https://app.daily.dev/feed-by-ids?id=6uX3MWeQL&amp;id=LtvttiElC" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The AI productivity slowdown is real</strong>: While marketing departments scream about 70% time savings, actual studies show developers are experiencing a 19% overall slowdown and a 50% increase in debugging time. We’ve traded the ‘boring’ work of writing code for the ‘miserable’ work of being a full-time human linter for probabilistic garbage. If your ‘productivity tool’ makes you spend half your day fixing its hallucinations, you’re not an engineer; you’re an unpaid intern for an LLM. <a href="https://app.daily.dev/feed-by-ids?id=Lb19kFYlj&amp;id=cY9Cx1sh3" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
