<h2>Must know</h2>
<ul>
<li><strong>Chrome zero-day under active exploitation</strong>: Google has released an urgent update for Chrome (version 145.0.7632.75/76) to patch CVE-2026-2441, a high-severity use-after-free vulnerability in its CSS font feature parsing that is actively being exploited in the wild. This is Chrome’s first zero-day of the year, and it could allow arbitrary code execution within the browser sandbox. The fix involves replacing reference-based iteration with a copy to prevent dangling pointers, a class of bug that Rust’s ownership model would prevent, as noted by sources like “they hacked CSS.” <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 has disclosed critical vulnerabilities in four widely used VS Code extensions—Live Server, Code Runner, Markdown Preview Enhanced, and Microsoft’s Live Preview—totaling 128 million installs. These flaws could enable file theft, remote code execution, and local network reconnaissance. While Microsoft silently patched Live Preview, three other maintainers have not responded, leaving millions of developers exposed, according to OX Security. <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 security bug (CW1226324) in Copilot Chat that allowed the AI to bypass confidentiality labels and Data Loss Prevention (DLP) policies, summarizing emails from users’ Sent Items and Drafts folders without authorization. The flaw, active since late January, highlights broader enterprise anxiety about AI tools accessing sensitive corporate data, with a fix rolling out in early February, as reported by BleepingComputer. <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>AI coding tools reshape development, but create new bottlenecks</strong>: While AI coding tools are projected to reach 90% adoption by enterprise developers by 2028 and increase pull request merges by 98%, they also create delivery bottlenecks at review, validation, and integration, leading to a widening gap between code generation and code shipping. <a href="https://app.daily.dev/feed-by-ids?id=alUW9sJXE&amp;id=guYQJk0m5&amp;id=okTNHJRhB&amp;id=sc19eWuhU&amp;id=vbdTHZjVN&amp;id=xQwTKeb7K&amp;id=ySpIGlBP7&amp;id=ygeEuk5Vz&amp;id=7hHra6Ai0&amp;id=WUJRo6v7h" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks launches custom agents for production AI</strong>: Databricks has launched Custom Agents (formerly Agent Framework), enabling developers to build, test, and deploy production-quality AI agents as fully managed Databricks Apps on serverless compute, complete with CI/CD integration and built-in memory via Lakebase. <a href="https://app.daily.dev/posts/QJn2roYcs" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA Run:ai boosts GPU fractioning for LLM inference</strong>: NVIDIA and Nebius AI Cloud benchmarks show that GPU fractioning with NVIDIA Run:ai can deliver 77% of full-GPU token throughput and 86% of concurrent user capacity with 0.5 GPU fractions, significantly maximizing GPU ROI for LLM inference at scale. <a href="https://app.daily.dev/posts/LsiRfCVWP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tailscale peer relays now generally available</strong>: Tailscale has made Peer Relays generally available, allowing users to deploy high-throughput relay nodes on any Tailscale device to improve network connectivity and performance, supporting static endpoints and enhanced observability. <a href="https://app.daily.dev/posts/Q0uzlfbeJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New developer tools for cloud deployment and streaming</strong>: Deno Deploy introduced <code>deno sandbox</code> for instant Linux microVM provisioning, TinyIce offers a lightweight, self-contained Icecast2-compatible audio streaming server, and Konta brings GitOps workflows to Docker Compose on low-resource VPS environments. <a href="https://app.daily.dev/feed-by-ids?id=7Ar3LQ3uM&amp;id=BHwbWnzAf&amp;id=QRFLXVuPF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel 12.52.0 adds makeMany() factory method</strong>: Laravel v12.52.0 introduces a <code>makeMany()</code> factory method for creating multiple unsaved model instances, alongside new helpers to skip lifecycle callbacks and atomic writes for the Blade compiler. <a href="https://app.daily.dev/posts/roJmm4d5R" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Red Hat launches enterprise Podman desktop</strong>: Red Hat has released an enterprise-supported build of Podman Desktop, addressing the demand from regulated industries for vendor-backed support for open-source container development tools, which has surpassed three million downloads. <a href="https://app.daily.dev/posts/iEOJ6oeIq" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI productivity gains are a myth, survey says</strong>: A 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 helping, but it’s definitely going to fire people. Makes perfect sense. <a href="https://app.daily.dev/posts/rfmYnNc4e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Claude built a C compiler from scratch, mostly</strong>: Anthropic released a marketing video claiming Claude built a C compiler ‘from scratch’ with ‘no human intervention’ that could compile Linux. Turns out, it used 37 years of GCC test suites, an online GCC oracle, required human restarts, cost $20,000 in API calls, and couldn’t actually boot Linux. But hey, 16 AI agents autonomously running for two weeks to produce a 100,000-line functional compiler is still pretty cool, if only they didn’t have to lie about it. <a href="https://app.daily.dev/posts/0csgEkRMy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Spotify CEO says senior devs don’t write code anymore</strong>: Spotify’s CEO claims senior developers haven’t written code since December 2025, instead supervising AI-generated code. Meanwhile, engineers report AI fatigue from reviewing large amounts of generated code, sometimes finding manual coding faster. It seems the ‘supervising’ part is just as much work, if not more, than actually writing the code, but with added existential dread. <a href="https://app.daily.dev/posts/7hHra6Ai0" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
