<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 value parsing. This is Chrome’s first zero-day of the year and is already being actively exploited in the wild. 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 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, collectively impacting over 128 million installations. 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 exposes 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. This flaw enabled Copilot to summarize emails from users’ Sent Items and Drafts folders without authorization, affecting Microsoft 365 business customers since late January. While a fix is rolling out, the incident highlights ongoing enterprise anxiety about AI tools accessing sensitive corporate data, a concern echoed by the European Parliament’s blocking of similar AI features. <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 management, security gateways, and evaluation frameworks. <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 offers a 1M token context window and is marketed as cheaper, but independent benchmarks suggest it’s a ‘token muncher’ due to adaptive thinking, while India and Japan push for sovereign AI 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 haven’t written code since December 2025, supervising AI instead, while others warn of AI fatigue and the need for developers to shift from coding to system design 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>Linux 7.0 performance regressions on Intel Panther Lake</strong>: Early benchmarks indicate performance regressions for Linux 7.0 on Intel Core Ultra Series 3 ‘Panther Lake’ hardware, affecting both CPU and iGPU performance compared to Linux 6.19 stable. <a href="https://app.daily.dev/posts/5CIXyHDbR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tailscale peer relays general availability</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, 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 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, 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>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>Anthropic’s C compiler claims are a bit of a stretch</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 even boot Linux due to a 16-bit x86 code size limitation. So, ‘from scratch’ means ‘with a lot of help and a huge budget,’ and ‘no human intervention’ means ‘except for the humans restarting it.’ Impressive, but maybe dial back the marketing hype a smidge. <a href="https://app.daily.dev/posts/0csgEkRMy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI productivity gains are still a myth for most executives</strong>: An NBER survey of nearly 6,000 executives found that over 80% reported no measurable impact from AI on productivity or employment over the past three years, despite 69% of businesses using some form of AI. Executives <em>anticipate</em> AI will eliminate 1.75 million jobs and deliver 1.4% productivity gains by 2028. So, we’re still in the ‘trust me, bro, it’s coming’ phase, while the actual numbers show a widening gap between AI investment hype and real-world outcomes. Maybe those 1.75 million jobs will be replaced by AI-powered surveys that also report no productivity gains. <a href="https://app.daily.dev/posts/rfmYnNc4e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Mandatory AI tooling metrics: The new ‘lines of code’</strong>: A developer at an unnamed company is critiquing a new internal metric mandating 80% of epics use AI tooling, arguing it measures tool usage rather than actual business outcomes. This is the classic ‘lines of code’ metric reborn, but now for AI. Because nothing says ‘innovation’ like forcing developers to use a tool, regardless of whether it actually helps. I’m sure the next metric will be ‘number of AI-generated commit messages per sprint,’ truly driving that business value. <a href="https://app.daily.dev/posts/8WhnSQAtX" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
