<h2>Must know</h2>
<ul>
<li><strong>NVIDIA donates GPU drivers to CNCF for cloud native AI</strong>: At KubeCon Europe, NVIDIA transferred ownership of its Dynamic Resource Allocation (DRA) Driver to the CNCF, enabling smarter GPU sharing and multi-node NVLink support directly within Kubernetes. This move, backed by AWS and Google Cloud, aims to standardize how heterogeneous accelerators are orchestrated for massive AI workloads. NVIDIA also introduced Grove, a new open-source API for AI workload orchestration, signaling a shift toward a more open infrastructure layer for LLM training and inference. <a href="https://app.daily.dev/posts/AoVkebmT8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic adds auto mode to Claude Code</strong>: Anthropic has introduced a new permissions system for its CLI tool, Claude Code, called ‘auto mode’ which uses a classifier to vet tool calls before execution. It automatically approves safe actions while blocking destructive operations like mass file deletions, escalating to a human only when it hits a block. This research preview for Team plan users aims to bridge the gap between tedious manual approvals and the risky skip-permissions flag during multi-file refactors. <a href="https://app.daily.dev/posts/Hgz3is1rZ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Supply chain attack hits LiteLLM via GitHub Actions</strong>: A misconfigured CI/CD pipeline in the Trivy project allowed attackers to exfiltrate a PyPI publishing token, leading to the compromise of LiteLLM versions v1.82.7 and v1.82.8. The malicious versions contained credential-stealing malware before being pulled from the registry. This incident highlights a growing trend of threat actors like TeamPCP systematically targeting the very open-source security tools developers trust to protect their infrastructure. <a href="https://app.daily.dev/feed-by-ids?id=bf7GKaXaH&amp;id=xiExzcA7d" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Cloud native community hits 20 million developers</strong>: The CNCF reports that 39% of all developers globally are now cloud native, with over half of backend engineers working within standardized platform engineering environments. <a href="https://app.daily.dev/feed-by-ids?id=6gZyGZFiz&amp;id=TnESiL7MY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Module Federation 2.0 migration at scale</strong>: Swiggy successfully unified its B2B dashboards by moving from S3-based micro-frontends to Module Federation, eliminating 638KB of duplicate JS per app and solving complex auth persistence issues. <a href="https://app.daily.dev/posts/0IigpOvLI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>TLA+ mental models for distributed systems</strong>: New insights from AWS and MongoDB engineers emphasize using TLA+ for state-centric reasoning and ‘slow is fast’ protocol coordination rather than just patching corner cases. <a href="https://app.daily.dev/posts/phcUUENWO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Running AI inference directly in the browser</strong>: QCon London highlights show WebGPU and WebNN are making local LLM inference viable for privacy-first apps, using quantization to shrink 7GB models down to 2GB. <a href="https://app.daily.dev/posts/FmvIxbZwO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Epic Games cuts 20 percent of workforce</strong>: CEO Tim Sweeney announced layoffs of over 1,000 employees, citing declining Fortnite revenue and high legal costs while pivoting focus toward Unreal Engine 6 development. <a href="https://app.daily.dev/posts/G1SUSytCF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>DSPy adoption remains low despite solving RAG pain</strong>: Engineers are reportedly reinventing DSPy’s typed I/O and prompt management patterns from scratch rather than adopting the framework’s software engineering abstractions for LLMs. <a href="https://app.daily.dev/posts/KkRBBl10a" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Vercel acquires new.website to boost v0 agents</strong>: The acquisition aims to integrate production-ready primitives like forms and databases directly into Vercel’s v0 generative UI platform. <a href="https://app.daily.dev/posts/TXo5OfLTR" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>The fight AI with AI marketing trap</strong>: RSAC 2026 is currently a sea of vendors claiming their ‘AI-powered’ security tools can stop ‘AI-powered’ attacks, a narrative ThreatLocker’s CEO calls ‘pure garbage.’ Most of these tools are just wrappers around the same LLMs that fail to determine software intent, yet they’re being sold as autonomous silver bullets for your SOC. <a href="https://app.daily.dev/feed-by-ids?id=f6diiN9F1&amp;id=k3pmCJTXy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Fantasy FAANGball and the productivity metric circus</strong>: We’ve reached peak industry absurdity where satirical sketches about ‘drafting’ engineers based on PR counts are indistinguishable from actual enterprise ‘developer velocity’ dashboards. If you’re evaluating your senior architects like they’re fantasy football players, you’ve officially traded engineering culture for a spreadsheet hallucination. <a href="https://app.daily.dev/posts/ngcGAMnP5" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The semantic layer will save us from SQL</strong>: MotherDuck and others are predicting a ‘text-to-SQL scandal’ in 2026 while simultaneously claiming humans will soon stop writing SQL entirely. It’s the classic hype cycle: admit the tech is currently unreliable and prone to disaster, then immediately pivot to saying it will replace every skilled professional in the stack by next Tuesday. <a href="https://app.daily.dev/feed-by-ids?id=X1JY9WQX8&amp;id=GLcGYCLOM" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
