<h2>Must know</h2>
<ul>
<li><strong>NVIDIA GTC 2026 unveils Vera CPU and DLSS 5</strong>: NVIDIA introduced 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 was unveiled, moving beyond simple upscaling to a neural rendering model that generates lighting and materials in real-time. While Digital Foundry noted impressive material realism, early demos faced criticism for an over-processed look and altered character faces in titles like Resident Evil: Requiem. <a href="https://app.daily.dev/feed-by-ids?id=te2HRAtVx&amp;id=5RfcDxaY5&amp;id=VcgVkME6V" 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 identified by Socket researchers. The campaign uses a bait-and-switch tactic where clean extensions later update to include hidden dependencies containing the GlassWorm loader. These extensions impersonate popular tools like ESLint and Prettier to silently deliver malware via automatic dependency installation. <a href="https://app.daily.dev/posts/yKzLDg3bN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Agentic AI governance and the rise of Minions</strong>: Stripe is now deploying over 1,300 automated coding agents, dubbed Minions, per week to handle tasks without human supervision. However, JFrog and Gartner warn of significant supply chain risks and permission sprawl, leading to the emergence of Agent Skills Registries for provenance. Research from Anthropic suggests shifting from role-based agent splitting to context-centric decomposition to manage the increasing complexity of these multi-agent systems. <a href="https://app.daily.dev/feed-by-ids?id=0yEYSAbXJ&amp;id=2DxOcZ3Ul&amp;id=MiuRs7TKU" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Linux 7.1 optimizes task scheduling and drops legacy UDP-Lite</strong>: The upcoming kernel improves scheduling latency via the SCX_ENQ_IMMED flag and removes unused UDP-Lite support for a 10% performance boost. <a href="https://app.daily.dev/feed-by-ids?id=9opkLYM2k&amp;id=BtK9i5ZUn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Amazon Timestream for InfluxDB 3 scales to 15 nodes</strong>: The new enterprise tier supports millions of data points per second with dedicated reader nodes and workload isolation for IoT and observability. <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 on-device hardware</strong>: Developed at Stanford, this framework prioritizes energy and latency constraints, handling nearly 89% of queries locally without cloud round-trips. <a href="https://app.daily.dev/posts/UjBsvfRxN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Oxyde ORM combines Pydantic with a Rust-powered core</strong>: This new asynchronous Python ORM outperforms SQLAlchemy and Tortoise in benchmarks while maintaining a familiar Django-style API. <a href="https://app.daily.dev/posts/GbDlMT2wK" 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 initial testers report rough edges with the sidebar and history features. <a href="https://app.daily.dev/posts/rVsK1uTPO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Karpathy releases tool to score AI exposure across occupations</strong>: The open-source tool uses Playwright and Gemini Flash to visualize BLS data and calculate digital AI exposure scores for 342 US jobs. <a href="https://app.daily.dev/posts/0kckLGbhk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>JetBrains sunsets Code With Me collaborative coding tool</strong>: The feature will be unbundled into a standalone plugin in version 2026.1 before public relay infrastructure is fully deactivated in 2027. <a href="https://app.daily.dev/posts/3UKmL0NfO" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The agentic AI bubble is definitely not a bubble</strong>: Ben Thompson and various VCs are working overtime to convince us that functional AI agents mean we’ve reached an inflection point, not a peak. They point to ‘Minions’ and ‘o1 reasoning’ as proof that enterprise investment is finally justified, conveniently ignoring that 96% of organizations report costs are higher than expected. It’s not a bubble if you just rename the runaway costs as ‘strategic infrastructure investment.’ <a href="https://app.daily.dev/feed-by-ids?id=39qO8ONeG&amp;id=c9X27okCy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>DLSS 5 is photorealism except for the faces</strong>: NVIDIA claims DLSS 5 is ‘semantics-aware’ and totally distinct from generative AI, right before Jensen Huang describes it as a fusion of graphics and generative models. Digital Foundry caught the tech ‘dramatically altering’ faces in Resident Evil, effectively crossing the uncanny valley and landing face-first in a vat of AI-filtered slop. Apparently, artistic intent is just another variable for the neural model to ‘predict’ away. <a href="https://app.daily.dev/posts/te2HRAtVx" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The xEngineer will see you now</strong>: Wix Engineering is pushing the ‘xEngineer’ concept, where your job shifts from writing code to being an ‘AI orchestrator’ and ‘problem definer.’ It’s a fancy way of saying you’ll spend your day debugging non-deterministic hallucinations instead of mastering syntax. We’re sure the ‘judgment and architecture’ phase will go great once the AI-written SQLite rewrite runs 2,000x slower than the original. <a href="https://app.daily.dev/feed-by-ids?id=VhO7qnQpo&amp;id=9pRi5GnN3" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
