<h2>Must know</h2>
<ul>
<li><strong>Chrome zero-day under active exploitation</strong>: Google just dropped an urgent Chrome update to patch CVE-2026-2441, a high-severity use-after-free bug in CSS font feature parsing that’s already being actively exploited. This is Chrome’s first zero-day of the year, and it could lead to arbitrary code execution inside the browser sandbox. The fix replaces reference-based iteration with a copy to prevent dangling pointers, a class of bug Rust’s ownership model would prevent. Update your Chrome, like, yesterday. <a href="https://app.daily.dev/posts/ISqaRw1RF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical vulnerabilities in popular VS Code extensions</strong>: OX Security found critical flaws in four popular VS Code extensions—Live Server, Code Runner, Markdown Preview Enhanced, and Microsoft’s Live Preview—totaling 128 million installs. These bugs allow file theft, remote code execution, and local network reconnaissance. Microsoft silently patched Live Preview, but most maintainers haven’t even responded. This highlights a massive supply chain risk in developer tooling, especially with AI-powered IDEs also affected. <a href="https://app.daily.dev/posts/IuDICegHN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Copilot chat security flaw exposed confidential emails</strong>: Microsoft confirmed a bug in Copilot Chat (CW1226324) that allowed the AI to bypass confidentiality labels and Data Loss Prevention (DLP) policies, summarizing emails from users’ Sent Items and Drafts without authorization. This flaw was active since late January, affecting Microsoft 365 business customers. It’s another reminder that AI tools accessing sensitive corporate data are a huge enterprise anxiety, especially after previous Copilot security concerns and the European Parliament blocking AI features on work devices. <a href="https://app.daily.dev/posts/jSmIdhAQU" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Linux 7.0 performance regressions on Intel panther lake</strong>: Early benchmarks show Linux 7.0 has performance regressions on Intel Core Ultra Series 3 ‘Panther Lake’ hardware, affecting both CPU and iGPU performance, despite expectations for Intel Xe driver improvements. <a href="https://app.daily.dev/posts/5CIXyHDbR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Asahi Linux progress on Apple M3 and Apple silicon support</strong>: Asahi Linux now has key peripherals like keyboard, touchpad, WiFi, NVMe, and USB3 working on Apple M3, with DisplayPort Alt Mode via USB-C also progressing in experimental branches after years of reverse engineering. <a href="https://app.daily.dev/posts/9Q140qigi" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel 7.0 developments and open source ecosystem updates</strong>: Linux kernel 7.0 is getting LoongArch architecture updates, removing the IBM Mwave modem driver, and FreeBSD is preparing a KDE Plasma desktop install option to simplify deployments. <a href="https://app.daily.dev/feed-by-ids?id=1hfw6pdps&amp;id=eVcCrU0bv&amp;id=nmU7F9kzS&amp;id=p4KzkvymW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Red Hat launches enterprise podman desktop</strong>: Red Hat released an enterprise-supported build of Podman Desktop, addressing the demand from regulated industries for vendor-backed support for open-source container development tools. <a href="https://app.daily.dev/posts/iEOJ6oeIq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA run:ai GPU fractioning for LLM inference performance</strong>: NVIDIA and Nebius AI Cloud benchmarked GPU fractioning with Run:ai, finding that 0.5 GPU fractions deliver 77% of full-GPU throughput and 86% of concurrent user capacity for LLM inference, maximizing GPU ROI. <a href="https://app.daily.dev/posts/LsiRfCVWP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tailscale peer relays general availability</strong>: Tailscale announced general availability of Peer Relays, allowing users to deploy high-throughput relay nodes on any Tailscale device to improve network connectivity and performance, even in restrictive cloud environments. <a href="https://app.daily.dev/posts/Q0uzlfbeJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tailscale integrates with huntress and fleet for automated device posture enforcement</strong>: Tailscale now integrates with Huntress EDR and Fleet MDM, allowing device posture and management state to be used in Tailscale access policies, automatically restricting access for non-compliant devices. <a href="https://app.daily.dev/feed-by-ids?id=L7iJSIIDS&amp;id=VukD3HNtV" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI productivity impact: all hype, no gains (yet)</strong>: An NBER survey of nearly 6,000 executives found that over 80% of firms reported no measurable impact from AI on productivity or employment over the past three years. Yet, these same executives anticipate AI will eliminate 1.75 million jobs by 2028. So, AI isn’t making us more productive, but it’s definitely going to take our jobs? Sounds like a classic case of ‘AI will solve all our problems, just not yet, and also it’s coming for you.’ <a href="https://app.daily.dev/posts/rfmYnNc4e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s ‘no human intervention’ C compiler was a $20k human-babysat clusterf</strong>k**: Anthropic’s marketing video claimed Claude built a C compiler ‘from scratch’ with ‘no human intervention.’ Turns out, it used 37 years of GCC test suites, an online GCC oracle, required human restarts, and cost $20,000 in API calls over two weeks with 16 parallel agents. And it couldn’t even boot Linux. The actual achievement of 16 AI agents autonomously running for two weeks to produce a 100,000-line functional compiler is impressive, but why ruin it with such dishonest marketing? Just tell us what it actually did, not what you wish it did. <a href="https://app.daily.dev/posts/0csgEkRMy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google launches lyria 3 AI music generation in gemini</strong>: Google just launched Lyria 3, an AI music generation model, integrated into Gemini. Now you can generate 30-second songs with lyrics from text, photos, or videos. Google claims it avoids mimicking specific artists and includes copyright filters, but ‘acknowledges these may not catch everything.’ So, get ready for a flood of AI-generated musical ‘slop’ that might or might not infringe on someone’s copyright, all watermarked with SynthID, because nothing says ‘art’ like a digital watermark. <a href="https://app.daily.dev/posts/eN6VW0Hvt" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
