<h2>Must know</h2>
<ul>
<li><strong>AI agents are reshaping infrastructure management</strong>: Red Hat’s Lightspeed MCP and Redpanda’s Agentic Data Plane are emerging to enable natural language-driven infrastructure health management and automated remediation, including generating Ansible Playbooks. These platforms aim to unify access layers, enforce policies, and provide granular observability for AI agents interacting with enterprise data systems. This shift promises to streamline DevOps tasks and enhance security in complex cloud environments. <a href="https://app.daily.dev/feed-by-ids?id=CoCDQAhlH&amp;id=cyEsB6IQv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kubernetes is becoming the OS for AI workloads</strong>: Kubernetes is solidifying its role as the operating system for AI/ML workloads, with new features in v1.35 like workload-aware scheduling and Dynamic Resource Allocation for GPUs. Cisco is also integrating eBPF-based Cilium and Tetragon into its enterprise networking to support high-throughput, low-latency AI workloads on Kubernetes. This evolution means Kubernetes is absorbing more AI infrastructure complexity, making it central to deploying and managing AI at scale. <a href="https://app.daily.dev/feed-by-ids?id=mAyDHBOpw&amp;id=DwIQAt8eB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloud outages are becoming the new normal</strong>: Cloud outages are increasing due to staffing challenges, enterprise complacency, and the growing complexity of hyperscale platforms. A recent 10-hour Azure outage, caused by a storage policy misconfiguration, highlights how human error in complex systems can lead to widespread disruption. Both cloud providers and enterprises need to invest more in operational excellence, training, architectural redundancy, and multicloud strategies to mitigate these recurring failures. <a href="https://app.daily.dev/posts/V2F1hpMgh" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI is changing cloud career paths</strong>: While AWS cloud skills remain valuable, they must now be paired with AI knowledge and a strong personal brand, as AI agents are increasingly displacing roles like solutions architect. <a href="https://app.daily.dev/posts/8yc6ZDNd6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>IaC’s real value is correctness, not just speed</strong>: Infrastructure-as-Code (IaC) is crucial for ensuring correctness by defining infrastructure as a single source of truth, preventing subtle mismatches between environments that cause unexpected failures. <a href="https://app.daily.dev/posts/wyL1mwgEd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New tools for AI-assisted infrastructure development</strong>: Platform Engineering Labs updated formae, an open-source IaC platform, with multi-cloud support and a Plugin SDK designed for AI-assisted development, aiming to avoid fragile state files. <a href="https://app.daily.dev/posts/oAyNaPkCA" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Continuous Delivery is key to managing AI-generated code</strong>: Continuous Delivery (CD) is essential for managing the increased code volume from AI, enabling teams to deploy software at any time through automation and fast feedback loops, reducing risk and pain. <a href="https://app.daily.dev/posts/r90VhIzh6" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Non-human identities are a critical attack surface</strong>: Non-Human Identities (NHIs) like tokens and API keys are a critical but often overlooked attack surface in cloud environments, requiring full lifecycle management for security and compliance. <a href="https://app.daily.dev/posts/X5JT6uroM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GPU fractioning optimizes AI inference costs</strong>: NVIDIA Run:ai’s GPU fractioning allows multiple LLM inference workloads to share GPUs efficiently, delivering significant token throughput and concurrent user capacity with minimal latency, maximizing GPU ROI. <a href="https://app.daily.dev/posts/LsiRfCVWP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Datadog Sheets centralizes security reporting</strong>: Datadog Sheets provides a spreadsheet-style interface to centralize, filter, and export vulnerability and compliance data from across code, infrastructure, and cloud resources, streamlining audit-ready reports. <a href="https://app.daily.dev/posts/zptmCv07k" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-generated Dockerfiles are a security risk</strong>: Apparently, AI-generated Dockerfiles are a thing now, and they come with ‘security and performance issues.’ Who would’ve thought that letting an AI churn out infrastructure code without human oversight could lead to problems? It’s almost like you need to understand what you’re doing, even when an AI is ‘helping.’ <a href="https://app.daily.dev/posts/Qwc1UQgC2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding agents are killing open source</strong>: Open-source maintainers are now ‘increasingly closing their projects’ because they’re ‘drowning in a flood of low-quality AI-generated submissions.’ So, AI is so good it’s making open source unusable. Daniel Stenberg even shut down cURL’s bug bounty because 95% of AI submissions were garbage. Maybe AI isn’t quite ready to take over the world, or at least not the world of pull requests. <a href="https://app.daily.dev/posts/BL3FTkDC5" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is making developers dumber</strong>: A software engineer argues that AI tools are shifting developer psychology from ‘imposter syndrome’ to the ‘Dunning-Kruger effect,’ making them feel like experts without real understanding. Apparently, the ‘productive struggle’ of debugging is now being skipped, leading to a generation of prompt-dependent coders. Who needs to learn fundamentals when an AI can just give you the answer, right? <a href="https://app.daily.dev/posts/8qhCrJSyB" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
