<h2>Must know</h2>
<ul>
<li><strong>Chrome zero-day under active exploitation</strong>: Google has released an urgent update for Chrome (145.0.7632.75/76 for Windows/Mac and 144.0.7559.75 for Linux) to patch CVE-2026-2441, a high-severity use-after-free vulnerability in its CSS font feature value map parsing. This is Chrome’s first zero-day of the year and is already being actively exploited. The bug, written in C++, can lead to arbitrary code execution within the browser sandbox, making immediate updates critical for all users. <a href="https://app.daily.dev/posts/ISqaRw1RF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical vulnerabilities found in popular VS Code extensions</strong>: OX Security disclosed critical vulnerabilities in four widely used VS Code extensions: Live Server, Code Runner, Markdown Preview Enhanced, and Microsoft’s Live Preview, collectively impacting over 128 million installs. These flaws enable file theft, remote code execution, and local network reconnaissance. While Microsoft silently patched Live Preview, most other maintainers have not responded, leaving millions of developers exposed. Developers are advised to update Live Preview and disable unused extensions. <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 ongoing enterprise anxiety about AI tools accessing sensitive corporate data. A fix is rolling out, but the incident underscores the critical need for robust security in AI integrations. <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 agent development, security, and governance accelerate</strong>: Enterprises are rapidly adopting and developing agentic AI, focusing on autonomous AI agents for various tasks, with new tools emerging for identity, security, and evaluation. <a href="https://app.daily.dev/feed-by-ids?id=2UmMyMAEf&amp;id=32FCqrqN0&amp;id=4fmmBjp20&amp;id=Ck3BiV0To&amp;id=bbSNghGeS&amp;id=fvUXvxfty&amp;id=hYwz99Mp8&amp;id=i8naoDZlv&amp;id=nicRNl9Cx&amp;id=pJKpsRETp" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI model performance, cost, and sovereign AI initiatives</strong>: Anthropic’s new Claude Sonnet 4.6 is cheaper but uses significantly more tokens, while India and Japan are making major strides in building sovereign AI infrastructure and models with NVIDIA. <a href="https://app.daily.dev/feed-by-ids?id=i4O6sQBIw&amp;id=j12nF8mSg&amp;id=jj2Kauaop&amp;id=lUz1VvbAr&amp;id=y0N7mzotz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI’s impact on software development and engineering roles</strong>: Spotify’s CEO claims senior developers are supervising AI-generated code, while experts debate AI’s role as an accelerator of existing strengths and weaknesses, emphasizing the need for strong fundamentals and outcome engineering. <a href="https://app.daily.dev/feed-by-ids?id=7hHra6Ai0&amp;id=WUJRo6v7h&amp;id=enHqyOYiB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tailscale peer relays general availability</strong>: Tailscale announced the general availability of Peer Relays, allowing users to deploy high-throughput relay nodes on any Tailscale device to improve network connectivity and performance, now 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>Databricks launches custom agents</strong>: Databricks launched Custom Agents (formerly Agent Framework) to enable developers to build, test, and deploy production-quality AI agents as fully managed Databricks Apps on serverless compute, with built-in memory and enterprise data access. <a href="https://app.daily.dev/posts/QJn2roYcs" 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 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 factory makeMany() method</strong>: Laravel v12.52.0 introduces a <code>makeMany()</code> factory method for creating multiple unsaved model instances, along with new helpers to skip lifecycle callbacks and improvements for <code>temporaryUploadUrl()</code> with local filesystems. <a href="https://app.daily.dev/posts/roJmm4d5R" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI productivity impact: all hype, no gains</strong>: A survey of nearly 6,000 executives found over 80% 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 now, but it’s definitely coming for your job later. Sounds like a classic case of ‘future promises, current excuses.’ <a href="https://app.daily.dev/posts/rfmYnNc4e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Claude C compiler: a marketing masterpiece of misinformation</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, and cost $20,000 in API calls. Oh, and it couldn’t actually boot Linux. But hey, ‘no human intervention’ sounds great on a press release, right? <a href="https://app.daily.dev/posts/0csgEkRMy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated code: the new ‘instant legacy’</strong>: Stack Overflow’s 2025 survey shows AI tool usage at 84% among developers, but trust dropped to 29%. Meanwhile, CircleCI reports a widening gap between code generation and code shipping, with main branch success rates at a five-year low. Apparently, AI makes code generation effortless, but it’s also accelerating the creation of ‘instant legacy code’ that’s hard to understand or debug. Who knew making things ‘effortless’ could be so much work? <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" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
