<h2>Must know</h2>
<ul>
<li><strong>Kubernetes ingress NGINX is getting archived</strong>: The Kubernetes Ingress NGINX project is being archived, meaning it will stop receiving security patches in March 2026. This affects roughly half of all cloud-native environments. Maintainer burnout and a lack of contributors are cited as reasons, with Kubernetes leadership pushing for a faster deprecation than maintainers desired. Teams are advised to migrate to alternatives like Cilium Ingress or Gateway API, Envoy Gateway, or Traefik, as reported by DEVCLASS and CNCF. This highlights the ongoing challenges of open-source sustainability for critical infrastructure. <a href="https://app.daily.dev/feed-by-ids?id=wK9Dxbori&amp;id=PptBbhnAu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Upwind security raises big bucks for runtime cloud security</strong>: Upwind Security just bagged a hefty $250 million in Series B funding, valuing them at $1.5 billion. They’re pushing an “inside-out” runtime cloud security platform that analyzes internal signals like network requests, contrasting with traditional agentless scanning. Despite initial hesitation from security teams about deploying internal agents, the company claims 900% year-over-year revenue growth and a doubled customer base, including Siemens and Peloton, as reported by TechCrunch. The funds are earmarked for AI security and developer-focused misconfiguration prevention. <a href="https://app.daily.dev/posts/LGgCBPA2D" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google cloud run gets NVIDIA RTX 6000 GPUs</strong>: Google Cloud Run now supports NVIDIA RTX PRO 6000 Blackwell GPUs in preview, allowing serverless deployment of large AI models (70B+ parameters) without the hassle of cluster management. These instances can scale from zero in under 5 seconds, offering 96GB vGPU memory and FP4/FP6 precision. This move aims to simplify real-time generative AI inference and model fine-tuning, with automatic driver installation and support for substantial vCPU and RAM configurations. <a href="https://app.daily.dev/posts/Vh4SzEb3w" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>India offers tax breaks to lure AI workloads</strong>: India is offering zero taxes through 2047 for foreign cloud providers running AI workloads from Indian data centers, provided services are sold outside the country, aiming to attract investment from tech giants. <a href="https://app.daily.dev/posts/UMOBKmYfu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open-source SwiftUI agent skill released</strong>: A new open-source SwiftUI Agent Skill has been released to help AI coding assistants build and refactor SwiftUI views, aiming to reduce technical debt by embedding SwiftUI expertise upfront. <a href="https://app.daily.dev/posts/LlEDXwb6h" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Claude AI integrates with Grav CMS for content management</strong>: A developer created an open-source MCP plugin for Grav CMS, enabling Claude AI to directly manage blog content via a JSON-RPC API with security features like authentication and rate limiting. <a href="https://app.daily.dev/posts/fBtDP8BjG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Wasmer beefs up Python and RISC-V support</strong>: Wasmer 7 introduces experimental async API support for Python and dynamic linking in WASIX, enabling native Python modules like Numpy to run in WebAssembly and improving RISC-V support. <a href="https://app.daily.dev/posts/fiQHrLvlC" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Astrological CPU scheduler for Linux</strong>: A new Linux kernel CPU scheduler, <code>scx_horoscope</code>, makes real scheduling decisions based on astrological principles, using planetary positions and zodiac signs to determine task priorities. <a href="https://app.daily.dev/posts/4O3zUvO4w" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 6.19-rc8 released</strong>: The Linux 6.19-rc8 kernel has been released, extending the development cycle by a week, with the stable 6.19 release expected soon. <a href="https://app.daily.dev/posts/WHsGYq8Jt" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vibe creates Linux VMs for LLM agents on ARM Macs</strong>: A new tool called Vibe creates zero-configuration Linux VMs on ARM Macs for sandboxing LLM agents, booting in ~10 seconds and providing better security than containers. <a href="https://app.daily.dev/posts/XJySsKWNc" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Microsoft scales back windows 11 AI overload</strong>: Microsoft is apparently “walking back” its AI overload in Windows 11, reviewing Copilot integrations and Windows Recall after “significant user backlash.” So, after aggressively shoving AI down everyone’s throats, they’re now pretending to listen. Don’t worry, the “under-the-hood AI efforts” will continue, because who needs user control when you have telemetry? <a href="https://app.daily.dev/posts/mR53bImT0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Tesla robotaxi crash rates are worse than human drivers</strong>: Tesla’s robotaxi fleet in Austin recorded 9 crashes over 500,000 miles, making its crash rate approximately 3x worse than human drivers for all crashes and 9x worse for police-reported incidents. This is despite having safety monitors in every vehicle. Meanwhile, Tesla continues to redact all crash narratives in its NHTSA reports, unlike Waymo, which operates truly driverless vehicles and has a better safety record. Transparency, who needs it? <a href="https://app.daily.dev/posts/M8WkrOIKM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>No-code platform aims to replace computer vision engineers</strong>: Anisoptera’s Dragonfly platform is here to replace those pesky computer vision engineers with a no-code solution, explicitly targeting the “AI talent bottleneck.” Because why hire expensive humans when you can just use a platform that promises to do it all? Engineers, better start brushing up on your “strategic decision-making” skills, because apparently, execution is for the machines now. <a href="https://app.daily.dev/posts/L9vFphH6F" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
