<h2>Must know</h2>
<ul>
<li><strong>Apple M5 chips prioritize AI performance amidst RAM shortages</strong>: Apple’s new M5-powered MacBook Air and Pro models claim up to 4x faster AI task performance than the M4, utilizing a new Fusion Architecture that merges two dies into a single SoC. Despite the performance gains, prices have increased by up to $400 due to a global RAM shortage driven by AI infrastructure demand. While the hardware supports Wi-Fi 7 and macOS Tahoe, Asahi Linux developers note that native Linux support for these chips is likely years away. <a href="https://app.daily.dev/posts/FGAdSpUhd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Critical security patches for Angular SSR and Django</strong>: Google has released urgent patches for Angular Server-Side Rendering to address a critical SSRF and header injection flaw that allows attackers to steal authorization headers or session cookies. Simultaneously, Django issued security updates (6.0.3, 5.2.12, 4.2.29) fixing a denial-of-service vulnerability in URLField and a race condition affecting file system permissions. Developers using these frameworks should prioritize these updates to mitigate remote code execution and data exfiltration risks. <a href="https://app.daily.dev/feed-by-ids?id=uyb72ug6C&amp;id=vB3tALZkk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>North Korean StegaBin campaign targets developer environments</strong>: The North Korean threat actor Famous Chollima is using a new campaign called StegaBin to distribute 26 malicious npm packages that typosquat popular libraries. These packages deploy a multi-stage infostealer targeting SSH keys, git credentials, and VSCode configurations across Linux, Windows, and macOS. The attack specifically targets Web3 and cryptocurrency developers, using Pastebin steganography to hide its command-and-control infrastructure. <a href="https://app.daily.dev/posts/QDsiIqiel" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Google Chrome shifts to a fortnightly release cycle starting September 2026</strong>: Chrome version 153 will move to a two-week cadence to ship performance fixes and web platform capabilities faster in response to AI browser competition. <a href="https://app.daily.dev/posts/SAMWZskhL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AWS launches open source plugins for AI agent cloud deployments</strong>: The new ‘deploy-on-aws’ plugin allows agents like Claude Code and Cursor to generate CDK code and estimate costs from natural language commands. <a href="https://app.daily.dev/posts/LBItm9B4M" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kubernetes v1.35 stabilizes in-place Pod resource resizing</strong>: This update allows CPU and memory updates on running containers without restarts, making Vertical Pod Autoscalers safe for stateful production workloads. <a href="https://app.daily.dev/posts/fjhms6Zq4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gemini 3.1 Flash-Lite targets high-volume developer workloads</strong>: Priced at $0.25 per million input tokens, this model is optimized for data processing speed rather than agentic orchestration. <a href="https://app.daily.dev/posts/z8znwLLdC" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust iterator abstraction causes 60x slowdown in Turbopuffer</strong>: The ‘zero-cost’ abstraction prevented auto-vectorization; switching to batched iterators dropped production query latency from 220ms to 47ms. <a href="https://app.daily.dev/posts/U3pwVYd72" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Helm 4 adopts WebAssembly for plugin isolation and portability</strong>: The new Wasm-based system replaces the Helm 3 model, offering 40% better latency and cross-architecture support for Kubernetes extensions. <a href="https://app.daily.dev/posts/W9pY3Qv5R" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel Cloud API and CLI reach general availability</strong>: The new REST API enables AI agents and CI/CD pipelines to programmatically manage deployments, databases, and resource hibernation. <a href="https://app.daily.dev/posts/BrFOlId9r" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>DevOps is dead according to the people selling AI agents</strong>: Cognition Lab claims human DevOps engineers are now ‘the exception’ because their Devin agent is supposedly everywhere. It’s a classic ‘buy our tool or become obsolete’ pitch that ignores the reality of managing complex, non-deterministic infrastructure. If DevOps were actually dead, we wouldn’t need so many humans to clean up the hallucinated YAML these agents produce. <a href="https://app.daily.dev/posts/LfBawqtwk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Two-tier valuations are the new unicorn makeup</strong>: Startups like Aaru and Serval are using ‘headline’ valuations to claim unicorn status while lead VCs actually buy in at less than half that price. This accounting gymnastics helps with recruiting and PR but creates a massive debt trap for the next funding round. It’s essentially a high-stakes game of pretend that mirrors the peak of the 2021 bubble. <a href="https://app.daily.dev/posts/cRCbCFvQ5" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>The AI information war is getting desperate</strong>: OpenAI and Anthropic are reportedly leaking selective financial projections to mask massive losses and inconsistent safety stances. While Anthropic plays the ‘ethical’ card by refusing some military deals, OpenAI is leaning into classified Pentagon networks for ‘all lawful use.’ It turns out ‘AI for humanity’ actually means ‘AI for whoever has the biggest government contract and the best PR firm.’ <a href="https://app.daily.dev/feed-by-ids?id=egxeYRWKg&amp;id=CiUt1uJYH" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
