<h2>Must know</h2>
<ul>
<li><strong>NVIDIA GTC 2026 reveals Vera CPU and DLSS 5</strong>: NVIDIA unveiled the Vera CPU, a custom 88-core processor designed specifically for agentic AI workloads with 1.2 TB/s memory bandwidth. On the consumer side, DLSS 5 moves beyond upscaling to full neural rendering, though early demos from Digital Foundry show controversial ‘AI-filtered’ artifacts on character faces. Jensen Huang also projected over $1 trillion in Blackwell and Vera Rubin chip orders through 2027, signaling no slowdown in infrastructure spend. <a href="https://app.daily.dev/feed-by-ids?id=te2HRAtVx&amp;id=5RfcDxaY5&amp;id=GlaCYECtS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GlassWorm malware hits Open VSX registry</strong>: A sophisticated supply-chain attack has compromised the Open VSX extension registry with at least 72 malicious extensions impersonating tools like ESLint and Prettier. According to Socket researchers, the malware uses a ‘clean-then-poison’ tactic where extensions update to include hidden dependencies that deliver the GlassWorm loader. Developers using VS Code with Open VSX should immediately audit their installed extensions and monitor for unauthorized dependency updates. <a href="https://app.daily.dev/posts/yKzLDg3bN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 7.1 optimizes scheduling and drops UDP-Lite</strong>: The upcoming Linux 7.1 kernel will include the SCX_ENQ_IMMED flag to significantly reduce task scheduling latency in sched_ext. The kernel is also officially removing UDP-Lite support due to lack of use and a persistent null-pointer bug, which surprisingly yielded a 10% performance boost in standard UDP tests on AMD EPYC hardware. Additionally, new HID drivers will finally bring full haptic and RGB support to Lenovo Legion Go handhelds. <a href="https://app.daily.dev/feed-by-ids?id=9opkLYM2k&amp;id=E28WnNt8S&amp;id=U4BmQzRBA" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<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>TC39 advances Temporal API to Stage 4</strong>: The long-awaited immutable date and time API is officially part of the ECMAScript 2026 specification, replacing the broken Date object. <a href="https://app.daily.dev/posts/WsjtjDMl5" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Amazon Timestream for InfluxDB 3 scales to 15 nodes</strong>: The managed service now supports multi-node enterprise clusters with dedicated reader nodes and one-click migration from Core editions. <a href="https://app.daily.dev/posts/rnQxpm2DT" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenJarvis brings local-first AI agents to devices</strong>: Developed at Stanford, this framework treats energy and latency as first-class constraints for running agents entirely on local hardware. <a href="https://app.daily.dev/posts/UjBsvfRxN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>JetBrains sunsets Code With Me collaboration tool</strong>: The real-time collaborative coding feature will be unbundled into a plugin before being fully deactivated in early 2027. <a href="https://app.daily.dev/posts/3UKmL0NfO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oxyde ORM combines Pydantic with a Rust core</strong>: A new Python ORM offers a Django-style API but uses Rust for high-performance asynchronous database operations. <a href="https://app.daily.dev/posts/GbDlMT2wK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Zephyr RTOS marks 10 years with massive adoption</strong>: A new report shows 70% of North American organizations now use the open-source RTOS in commercial embedded products. <a href="https://app.daily.dev/posts/VibQU9lW9" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The $100,000 single agent session</strong>: While everyone is busy ‘agentizing’ their workflows, 96% of organizations report AI costs are higher than expected, with some single agent sessions hitting six figures. We’ve moved from ‘cloud sprawl’ to ‘zombie agent sprawl,’ where autonomous scripts burn through credits while humans are still trying to figure out what the agent was supposed to do. It’s a masterclass in turning venture capital into pure compute heat without a clear ROI. <a href="https://app.daily.dev/posts/c9X27okCy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated code is the new technical debt</strong>: A recent study on Cursor AI adoption found that while it boosts short-term velocity, it causes a persistent spike in static analysis warnings and code complexity. We are essentially using AI to write code 2,000x slower than human-written equivalents—like the recent SQLite Rust rewrite—just to say we used ‘agentic’ workflows. It’s not ‘xEngineering’; it’s just high-speed legacy code generation. <a href="https://app.daily.dev/feed-by-ids?id=9pRi5GnN3&amp;id=0yEYSAbXJ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Governance capture by the AI elite</strong>: Silicon Valley’s ‘AI elite’ are systematically converting political donations into regulatory influence to ensure their proprietary models get government contracts. By using geopolitical fear-mongering about China, they’ve managed to dismantle federal oversight while pushing for massive energy subsidies. It’s a classic case of the regulated becoming the regulator, all wrapped in a ‘national security’ flag. <a href="https://app.daily.dev/feed-by-ids?id=aVcUWoSUS&amp;id=ZsR4Q7LWZ" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
