<h2>Must know</h2>
<ul>
<li><strong>GlassWorm malware hits Open VSX extension registry</strong>: A sophisticated supply-chain attack has compromised the Open VSX registry with at least 72 malicious extensions impersonating tools like ESLint, Prettier, and Claude Code. According to Socket researchers, the malware uses a delayed execution tactic where clean extensions later update to pull in hidden dependencies containing the GlassWorm loader. Because VS Code often installs these dependencies silently, developers are urged to audit their extension updates and restrict installs to trusted publishers. <a href="https://app.daily.dev/posts/yKzLDg3bN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA GTC 2026 unveils Vera CPU and DLSS 5</strong>: NVIDIA has announced the Vera CPU, a custom 88-core processor designed specifically for agentic AI workloads with 1.2 TB/s memory bandwidth. On the graphics front, DLSS 5 was unveiled, moving beyond simple upscaling to a neural rendering model that generates lighting and materials from scratch. While Digital Foundry noted impressive realism in materials, early demos have faced criticism for an over-processed look and dramatically altered character faces. <a href="https://app.daily.dev/feed-by-ids?id=5RfcDxaY5&amp;id=te2HRAtVx&amp;id=VcgVkME6V" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vite 8.0 swaps Rollup for Rust-based Rolldown</strong>: The Vite team has officially released version 8.0, featuring Rolldown as the primary bundler to replace both esbuild and Rollup. Built on the Oxc framework in Rust, Rolldown reportedly delivers build speeds 10 to 30 times faster than its predecessors while maintaining compatibility. This release also introduces Wasm support in SSR and a new DevTools suite for deep build analysis, as reported by The Register. <a href="https://app.daily.dev/posts/mksK39aMS" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Amazon Timestream for InfluxDB 3 scales to 15 nodes</strong>: The new enterprise tier supports clusters with up to 13 dedicated reader nodes and 4 writer nodes for high-concurrency IoT workloads. <a href="https://app.daily.dev/posts/rnQxpm2DT" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 7.1 to drop UDP-Lite and optimize task scheduling</strong>: The upcoming kernel removes the disused UDP-Lite protocol for a 10% performance gain and adds the SCX_ENQ_IMMED flag to improve scheduling latency. <a href="https://app.daily.dev/feed-by-ids?id=9opkLYM2k&amp;id=TVQGiQ3e0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kagi Orion browser enters public beta for Linux</strong>: The WebKit-based browser is now available as a Flatpak, though it currently lacks extension support and Kagi Sync. <a href="https://app.daily.dev/posts/rVsK1uTPO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oxyde ORM brings Rust-powered performance to Python</strong>: This new asynchronous ORM uses a Rust core and Pydantic v2 to outperform SQLAlchemy and Tortoise in raw operations per second. <a href="https://app.daily.dev/posts/GbDlMT2wK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenJarvis framework targets local-first AI agents</strong>: Developed at Stanford, this framework provides shared primitives for on-device AI, treating energy and latency as primary constraints. <a href="https://app.daily.dev/posts/UjBsvfRxN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Pushduck simplifies universal file uploads across 16 frameworks</strong>: A type-safe TypeScript library that abstracts S3, R2, and Spaces storage while handling presigned URLs and chunked uploads. <a href="https://app.daily.dev/posts/xiGGOItVN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Pompelmi offers in-process malware scanning for Node.js</strong>: An open-source library providing YARA-based scanning and ZIP bomb protection without requiring external daemons or cloud APIs. <a href="https://app.daily.dev/posts/xJOXpL6zn" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The agentic AI cost explosion</strong>: While marketing decks promise AI agents will replace your entire engineering team, reality is hitting the balance sheet with 96% of organizations reporting costs higher than expected. Some enterprise agent sessions have reportedly hit $100,000, proving that ‘autonomous’ is often just a synonym for ‘unbounded API billing.’ <a href="https://app.daily.dev/posts/c9X27okCy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated code quality reckoning</strong>: Founders at Codestrap are predicting a massive ‘reckoning’ as unreviewed AI-slop begins to rot codebases from the inside out. They cite a recent SQLite rewrite in Rust by an AI that managed to be 2,000 times slower than the original, reminding us that speed to commit is not the same as speed to production. <a href="https://app.daily.dev/feed-by-ids?id=9pRi5GnN3&amp;id=yEdWFqSbz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA DLSS 5 is definitely not generative AI except when it is</strong>: NVIDIA CEO Jensen Huang spent GTC describing DLSS 5 as a fusion of graphics and generative AI, only for the official PR to immediately pivot and call it ‘distinct from generative AI.’ It seems the marketing department is terrified that if they admit it’s just a fancy deepfake for your GPU, people might notice the ‘dramatically altered’ faces in the demos. <a href="https://app.daily.dev/posts/te2HRAtVx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
