<h2>Must know</h2>
<ul>
<li><strong>AI coding assistants increase planning demands and security risks</strong>: AI coding assistants, while boosting code writing speed, demand significantly more upfront architectural planning and specification work. A 2025 study indicates developers using AI assistants took 19% longer to complete tasks overall, despite feeling faster. Code review times on high-AI-usage teams have surged by 91%, and a staggering 45% of AI-generated code contains security vulnerabilities, according to a report. This suggests that treating AI as a junior developer needing clear direction, rather than an autonomous agent, is crucial to avoid the ‘zig-zag problem’ of unproductive iterations and ‘context pollution’ during debugging. <a href="https://app.daily.dev/posts/pYjmmMIkn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Open source sustainability threatened by AI code generation</strong>: The open-source ecosystem faces an existential threat as AI-powered code generation, or ‘vibe coding,’ allows AI agents to assemble software without direct interaction with project maintainers. Tailwind CSS, for instance, reported an 80% drop in revenue and laid off 75% of its staff, despite 75 million monthly downloads, as developers bypass traditional documentation for AI tools. A research paper models a decline in OSS quality and availability due to reduced maintainer incentives, advocating for new compensation mechanisms to sustain this critical infrastructure, as reported by Frontend Masters. <a href="https://app.daily.dev/feed-by-ids?id=3KjVRqoNX&amp;id=YRC8HFFhU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft launches Maia 200 AI inference accelerator</strong>: Microsoft has unveiled the Maia 200, a custom AI inference accelerator built on TSMC’s 3nm process, featuring native FP8 and FP4 tensor cores, 216GB HBM3e memory, and 272MB on-chip SRAM. It boasts over 10 petaFLOPS in FP4 precision and claims 30% better performance per dollar than current hardware. Designed for GPT-5.2 models, Microsoft 365 Copilot, and synthetic data generation, the Maia 200 is initially deploying in Microsoft’s US Central datacenter, with plans for expansion and an SDK preview. <a href="https://app.daily.dev/posts/HBgJElvHn" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>React Server Components vulnerable to DoS attacks</strong>: Multiple high-severity denial-of-service vulnerabilities (CVE-2026-23864, CVSS 7.5) have been identified in React Server Components, affecting React 19.0.x through 19.2.x and Next.js versions 13.x through 16.x, allowing server crashes or excessive CPU usage via crafted HTTP requests. <a href="https://app.daily.dev/posts/igYAcLEa7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Node.js 25.5.0 enhances security and features</strong>: Node.js 25.5.0 upgrades root certificates to NSS 3.119, introduces LIEF for Single Executable Application (SEA) generation, adds an ‘ignore’ option to fs.watch, and defaults to SQLite defensive mode for enhanced security. <a href="https://app.daily.dev/posts/BrQRGrBKE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google launches optimized mobile/edge LLM for function calling</strong>: Google released FunctionGemma, a 270M parameter variant of the Gemma 3 model, specifically optimized for efficient function calling on mobile and edge hardware, translating natural language into structured API calls with 85% accuracy. <a href="https://app.daily.dev/posts/DFyxBQQFb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Docker Sandboxes isolates AI coding agents with microVMs</strong>: Docker Sandboxes now uses microVM-based isolation for AI coding agents like Claude Code and Gemini CLI, allowing them to install packages and modify files without affecting the host system, currently available on macOS with Windows and Linux support planned. <a href="https://app.daily.dev/posts/6YdOupFK7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI upgrades ChatGPT code execution with Bash and package support</strong>: OpenAI’s ChatGPT code execution feature now supports running Bash commands and installing packages via pip and npm through an internal proxy, significantly expanding its utility for all users, including free-tier. <a href="https://app.daily.dev/posts/7V1v7MpYn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ktor 3.4 enhances HTTP features and API documentation</strong>: Ktor 3.4 introduces duplex streaming for the OkHttp engine, enabling simultaneous HTTP/2 communication, and adds a new runtime API for dynamic endpoint documentation, including auto-generation of Swagger (OpenAPI) docs. <a href="https://app.daily.dev/posts/FLMxxYChP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Uncloud Kubernetes alternative gains momentum</strong>: Uncloud, an open-source deployment platform using Docker, Caddy, and WireGuard, is gaining traction among small teams as a simpler Kubernetes alternative, offering single-command deployments and nearing a stability milestone with plans for a managed control center. <a href="https://app.daily.dev/posts/FrS1fa85P" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Cursor AI agents build a browser ‘from scratch’ (spoiler: they didn’t)</strong>: Cursor’s CEO proudly announced their AI agents built a fully functional web browser ‘from scratch’ in a week, generating over 3 million lines of code. Independent developers, however, quickly found the browser barely compiles, doesn’t work, and heavily relies on existing projects like Servo and QuickJS. So, ‘from scratch’ apparently means ‘cobbled together with existing parts and still broken,’ but hey, the marketing spin was top-notch! <a href="https://app.daily.dev/posts/DovPZUpOP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is making software engineering all about SRE, apparently</strong>: Two identical articles claim AI makes code-writing so easy that the new ‘core value’ for software engineers is operational excellence and reliability. So, building a working demo is ‘straightforward’ now, but maintaining it is the real challenge. Apparently, Site Reliability Engineering (SRE) is the future, because who needs to write good code when AI can just churn it out, right? It’s almost like they’re saying AI is great for quantity, but humans are still stuck with the quality control. <a href="https://app.daily.dev/feed-by-ids?id=cIiWYuVU9&amp;id=mMOsDNB8r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding assistants: the energy footprint nobody talks about</strong>: Turns out, your AI coding agent isn’t just a token-guzzling genius; it’s also an energy hog. A median Claude Code session uses about 41 Wh, which is 138 times more than a typical chatbot query, according to Tech Lead Digest. Daily usage can hit 1,300 Wh, comparable to running an extra dishwasher cycle. So, while we’re all marveling at AI’s productivity, maybe we should also start thinking about its carbon footprint. Or, you know, just keep ignoring it like we do with everything else. <a href="https://app.daily.dev/posts/QxbydGQJy" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
