<h2>Must know</h2>
<ul>
<li><strong>Chrome zero-day under active exploitation</strong>: Google released an urgent Chrome update (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. The bug allows arbitrary code execution within the browser sandbox, making immediate updates critical. This marks Chrome’s first zero-day of 2026, highlighting the continuous threat landscape for web browsers. <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 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, the maintainers of the other three extensions 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 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, summarizing emails from users’ Sent Items and Drafts folders without authorization. The flaw, active since late January, highlights broader enterprise anxieties about AI tools accessing sensitive corporate data, especially given 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>NVIDIA and Sarvam AI optimize LLM inference</strong>: NVIDIA and Sarvam AI collaborated to achieve a 4x inference speedup for Sarvam’s Sovereign 30B multilingual MoE model, leveraging H100 and Blackwell GPUs with kernel-level optimizations and NVFP4 quantization. This co-design approach ensures strict voice-agent SLAs, demonstrating the importance of full-stack optimization for AI at scale. <a href="https://app.daily.dev/posts/yKP3YkWSF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks flexible node types now generally available</strong>: Databricks released Flexible Node Types, allowing automatic fallback to compatible instance types during cloud provider stockouts and optimizing Spot Instance usage across AWS, Azure, and GCP, preventing cluster launch failures and improving cost efficiency. <a href="https://app.daily.dev/posts/7IXsRxCOl" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks automates data optimization with Predictive Optimization</strong>: Databricks’ Predictive Optimization in Unity Catalog is now the default for all new managed tables, autonomously handling compaction, VACUUM, statistics, and clustering, leading to up to 22% faster queries and significant storage cost savings. <a href="https://app.daily.dev/posts/GSuspp2K8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel adds official Svelte starter kit</strong>: Laravel now offers an official Svelte starter kit, integrating Svelte 5, TypeScript, Inertia 2, Tailwind CSS, and shadcn-svelte, providing a robust foundation for new projects alongside existing React, Vue, and Livewire options. <a href="https://app.daily.dev/posts/runkXqODs" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Deno Deploy introduces instant Linux microVMs</strong>: Deno Deploy launched <code>deno sandbox</code>, a new subcommand for provisioning cloud-hosted Linux microVMs and SSH-connecting in under a second, complete with a secrets mechanism that injects values at the network level without exposing raw secrets. <a href="https://app.daily.dev/posts/7Ar3LQ3uM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Apache Flink Kubernetes Operator 1.14.0 released with Blue/Green deployments</strong>: Apache Flink Kubernetes Operator 1.14.0 introduces native Blue/Green deployment support, enabling zero-downtime upgrades for stateful streaming applications by managing parallel environments, automating savepoints, and switching traffic only after stability. <a href="https://app.daily.dev/posts/RfrUoGk5N" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Trivago scales Kafka sinks with KEDA</strong>: Trivago successfully used KEDA (Kubernetes Event-Driven Autoscaling) with its Kafka scaler to scale down idle Kafka sink services to zero replicas, reducing average daily consumption from 50 to 1-2 replica-hours by using consumer group lag as the scaling signal. <a href="https://app.daily.dev/posts/vltl2g6Wl" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI productivity impact: much ado about nothing?</strong>: A recent NBER survey of nearly 6,000 executives found that over 80% reported no measurable impact from AI on productivity or employment, with 89% seeing no gains in the last three years. Despite 69% of businesses using some form of AI, the hype continues to outpace actual, quantifiable business outcomes. Executives do, however, anticipate AI will eliminate 1.75 million jobs by 2028, which is a bold prediction given the current lack of impact. <a href="https://app.daily.dev/posts/rfmYnNc4e" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s C compiler: a marketing masterpiece or a technical marvel?</strong>: Anthropic claimed Claude built a C compiler ‘from scratch’ with ‘no human intervention,’ capable of compiling Linux and Doom. The reality, as detailed by critics, is that 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. It seems ‘from scratch’ now means ‘with a lot of help and a hefty bill,’ but hey, the marketing video was probably great. <a href="https://app.daily.dev/posts/0csgEkRMy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated code: the dumbest metric is back</strong>: Martin Fowler warns that AI is accelerating software development, but also risks becoming a ‘debt accelerator’ if foundational practices are ignored. Meanwhile, the metric of Lines of Code (LOC) is making a dangerous comeback due to AI code generation hype, despite being discredited decades ago. Apparently, AI makes motion effortless but leaves judgment unchanged, leading to ‘instant legacy code’ that nobody fully understands. What could possibly go wrong? <a href="https://app.daily.dev/posts/WUJRo6v7h" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
