<h2>Must know</h2>
<ul>
<li>AI coding assistants are still a mess, but getting better: Despite the hype, AI coding assistants like Claude Opus 4.5 and GPT-5.2 are still struggling with basic tasks, often making excuses instead of debugging, according to one developer’s critique. However, Docker Sandboxes is now offering microVM isolation for AI agents like Claude Code and Gemini CLI, allowing them to install packages and modify files without trashing your host system. Meanwhile, OpenAI has quietly upgraded ChatGPT’s code execution to include Bash commands and package installation, making it more powerful but also potentially more dangerous if not properly sandboxed. <a href="https://app.daily.dev/feed-by-ids?id=Iw3ITFPgH&amp;id=6YdOupFK7&amp;id=7V1v7MpYn" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Microsoft’s new AI chip and Windows dev tool: Microsoft just dropped the Maia 200, a custom AI inference accelerator built on TSMC’s 3nm process, boasting over 10 petaFLOPS in FP4 and claiming 30% better performance per dollar. This beast is slated for GPT-5.2 models and Microsoft 365 Copilot, with initial deployment in their US Central datacenter. In other news, Microsoft also launched ‘winapp,’ a CLI tool to simplify Windows app development outside Visual Studio, though developers are reportedly skeptical it will fix long-standing platform issues, according to DEVCLASS. <a href="https://app.daily.dev/feed-by-ids?id=HBgJElvHn&amp;id=xa7n8KICk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>EU targets Grok AI for harmful content: The EU Commission has launched an investigation into Elon Musk’s Grok AI over allegations it generated sexualized images of women and minors, potentially violating the Digital Services Act (DSA). This comes after the Center for Countering Digital Hate claimed Grok produced millions of such images. If violations are confirmed, X (Grok’s parent company) could face fines up to 6% of its global annual turnover, and Grok might be pulled from the EU market. Meanwhile, ChatGPT and Claude have been caught citing Grokipedia, an AI-generated encyclopedia known for inaccuracies and bias, raising questions about source validation in major AI platforms. <a href="https://app.daily.dev/feed-by-ids?id=tpCDWbga3&amp;id=KoLpeTHv0&amp;id=SDpas3mop" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li>AI is transforming incident response for SREs: NeuBird’s Hawkeye, backed by AWS and Microsoft, is using agentic AI to automate incident investigation for SREs, claiming over 90% reduction in mean time to recovery by autonomously querying telemetry and identifying root causes in read-only mode. <a href="https://app.daily.dev/feed-by-ids?id=cIiWYuVU9&amp;id=mMOsDNB8r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Nvidia invests big in AI compute capacity: Nvidia is pouring $2 billion into CoreWeave to accelerate the construction of over 5 gigawatts of AI computing infrastructure by 2030, integrating Nvidia’s Rubin chip architecture and other solutions, despite CoreWeave’s significant debt. <a href="https://app.daily.dev/posts/suyuzcb2U" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Linux kernel 6.19 rolls back scheduler changes: The Linux 6.19 kernel’s rc7 release candidate has disabled the NEXT_BUDDY feature for the EEVDF scheduler due to performance regressions in workloads like MySQL and SPECjbb, with benchmarks confirming the issues. <a href="https://app.daily.dev/feed-by-ids?id=85ivk605B&amp;id=mNRQJiXzS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Google launches mobile/edge LLM for function calling: Google released FunctionGemma, a 270M parameter variant of the Gemma 3 model, optimized for function calling on mobile and edge hardware, translating natural language into structured API calls with a notable accuracy increase from 58% to 85%. <a href="https://app.daily.dev/posts/DFyxBQQFb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Ambitious Rust-based language emerges with AI help: Rust core team member Steve Klabnik unveiled Rue, a new memory-safe systems programming language that ditches the borrow checker for ‘inout’ parameters, developed with substantial assistance from Anthropic’s Claude AI, which generated 70,000 lines of compiler code. <a href="https://app.daily.dev/posts/scnD6visU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Ambitious Rust-based language emerges with AI help: Steve Klabnik, a Rust core team member, has unveiled Rue, a new memory-safe systems programming language that removes the borrow checker in favor of ‘inout’ parameters, developed with substantial assistance from Anthropic’s Claude AI, which generated approximately 70,000 lines of compiler code. <a href="https://app.daily.dev/posts/scnD6visU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI-powered weather forecasting goes open source: Nvidia released three open-source AI weather forecasting models—Earth-2 Nowcasting, Earth-2 Medium Range, and Earth-2 Global Data Assimilation—aiming to democratize access to advanced forecasting and reduce reliance on government supercomputers. <a href="https://app.daily.dev/posts/AZxwXEsfz" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>New Docker Sandboxes for AI coding agents: Docker Sandboxes now offers microVM-based isolation for AI coding agents like Claude Code, Gemini CLI, Codex, and Kiro, allowing them to install packages and modify files without impacting 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>OpenTelemetry instrumentation still a challenge for AI: A benchmark called OTelBench found that even advanced AI models like Claude Opus 4.5 only achieved a 29% success rate in instrumenting codebases with OpenTelemetry, struggling with context propagation and failing entirely for some languages like Java and Swift. <a href="https://app.daily.dev/posts/mI3Z3Ohub" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>KDE Plasma 6.6 shifts to systemd, drops FreeBSD support: KDE Plasma 6.6 beta is moving to a systemd-dependent login manager, effectively ending FreeBSD support and signaling an increasing Linux-specific direction for the project, mirroring similar changes in GNOME. <a href="https://app.daily.dev/posts/satawafBY" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>New Node.js 25.5.0 release: Node.js 25.5.0 is out, featuring an upgrade to NSS 3.119 root certificates, a new <code>--build-sea</code> flag for Single Executable Application generation, and an <code>ignore</code> option for <code>fs.watch</code>, alongside various bug fixes and dependency updates. <a href="https://app.daily.dev/posts/BrQRGrBKE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Pandas 3 released with new features: Pandas 3 has been officially released and is available via pip and uv package managers, with new features and changes documented in a dedicated YouTube playlist, though specific details are not yet provided. <a href="https://app.daily.dev/posts/9EFNgACx5" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Ktor 3.4 enhances HTTP and API docs: Ktor 3.4 introduces duplex streaming for the OkHttp engine, enabling full-duplex 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>AWS updates EC2, ECR, and CloudWatch: AWS announced new EC2 G7e instances with NVIDIA Blackwell GPUs, promising 2.3x improved inference performance, and updated Amazon Elastic Container Registry (ECR) with cross-repository layer sharing for faster push operations and reduced storage costs. <a href="https://app.daily.dev/posts/lbS5HEqo8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Uncloud gains traction as Kubernetes alternative: Uncloud, an open-source deployment platform built on Docker, Caddy, and WireGuard, is gaining momentum among small teams as a simpler Kubernetes alternative, supporting single-command deployments from source to production without container registries. <a href="https://app.daily.dev/posts/FrS1fa85P" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AMDGPU Linux driver adds HDMI VRR and ALLM: AMDGPU Linux kernel driver patches now introduce HDMI Variable Rate Refresh (VRR) and Auto Low Latency Mode (ALLM) for gaming on HDMI displays, aiming for feature parity with FreeSync capabilities on DisplayPort and Windows. <a href="https://app.daily.dev/posts/bsF1QAXcQ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Laravel Boost v2.0 adds skills system: Laravel Boost v2.0 introduces a new Skills system for developers to extend and customize their workflow, featuring skill synchronization, overrides, and a <code>boost:add-skill</code> Artisan command to fetch skills from GitHub. <a href="https://app.daily.dev/posts/xZZMD5bdo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>New Laravel Herd Worktree tool for Claude Code: Claude Code has a new skill called Laravel Herd Worktree, which automates the setup of isolated development environments for Laravel projects, including auto-configuring settings, installing dependencies, and integrating Herd and Vite. <a href="https://app.daily.dev/posts/9nZ5eRVBM" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI-powered research agent released: StepFun AI released Step-DeepResearch, a 32-billion-parameter research agent model built on Qwen2.5, designed for long-horizon research tasks with capabilities like planning, deep information seeking, reflection, and report generation. <a href="https://app.daily.dev/posts/cOrYC12OX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>OpenSearch Kubernetes Operator 3.0 Alpha released: OpenSearch Kubernetes Operator 3.0 Alpha is out with major stability improvements, including quorum-safe rolling restarts, multi-namespace support, TLS certificate hot reloading, and gRPC API support, alongside over 100 other changes. <a href="https://app.daily.dev/posts/WZU2opC0j" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>NVIDIA TensorRT for RTX gets adaptive inference: NVIDIA’s new TensorRT for RTX introduces adaptive inference, allowing AI engines to automatically optimize performance during runtime with features like Dynamic Shapes Kernel Specialization, built-in CUDA Graphs, and runtime caching, promising a 1.32x speedup for the FLUX.1 AI model on RTX 5090. <a href="https://app.daily.dev/posts/pMSQ81agb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Vite releases Rust-based bundler Rolldown 1.0 RC: Vite has released Rolldown 1.0 RC, a Rust-based bundler for JavaScript applications designed to combine fast performance with compatibility for Rollup plugins and ecosystem, featuring built-in transforms for TypeScript and JSX. <a href="https://app.daily.dev/posts/q9ho7IfK8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Aspire 13.1 adds MCP integration: Aspire 13.1 has been released, featuring Model Context Protocol (MCP) integration to enable AI coding agents to detect and interface with active resources, alongside CLI updates, dashboard enhancements, and improved Azure deployment processes. <a href="https://app.daily.dev/posts/rWV2xAd5H" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><a href="http://llms.py" target="_blank" rel="noopener nofollow">llms.py</a> v3 released with new architecture: <a href="http://llms.py" target="_blank" rel="noopener nofollow">llms.py</a> v3 has been released, introducing an extensible architecture supporting over 530 models from 24 providers, a new plugin system, Gemini RAG integration, and native support for tool/function calling and MCP. <a href="https://app.daily.dev/posts/BIj2KZjMG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><a href="http://X.Org" target="_blank" rel="noopener nofollow">X.Org</a> libraries get maintenance updates: Multiple <a href="http://X.Org" target="_blank" rel="noopener nofollow">X.Org</a> libraries, including libXmu 1.3.0, libXext 1.3.7, and libxkbfile 1.2.0, have received maintenance updates in early 2026, as the <a href="http://X.Org" target="_blank" rel="noopener nofollow">X.Org</a> community awaits further developments on the <a href="http://X.Org" target="_blank" rel="noopener nofollow">X.Org</a> Server Git branch reorganization. <a href="https://app.daily.dev/posts/7NN9ETfne" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>KotlinConf 2026 announces first speakers: KotlinConf 2026 has announced its first round of speakers, described as experts influencing Kotlin’s development, with the full schedule and speaker list to be released later for the event that typically draws participants from over 50 countries. <a href="https://app.daily.dev/posts/7mtHzx4mm" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Firefox Nightly defaults to split-view mode: Firefox Nightly now has Split View mode enabled by default, allowing users to view two web pages side-by-side within a single browser window, though this feature is not yet in the main or release versions of Firefox. <a href="https://app.daily.dev/posts/tvHaqzfAK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>New web component library for animated backgrounds: Beautiful Backgrounds, a new TypeScript library, offers six lightweight and customizable animated web components like Digital Rain and Hexagon Wave for dynamic website backgrounds, with Storybook integration for real-time parameter adjustments. <a href="https://app.daily.dev/posts/qpkSB70xe" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Obsidian VimRC plugin commands overview: Obsidian’s VimRC plugin now offers over 400 commands for advanced customization across editing, workspace, file management, markdown operations, and navigation, all accessible through the developer console. <a href="https://app.daily.dev/posts/a7bLuMXND" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li>AI-generated browser built ‘from scratch’ by AI agents: Cursor’s CEO claimed their AI agents built a fully functional web browser ‘from scratch’ in a week, generating over 3 million lines of code. Independent developers, however, found the public GitHub repo barely compiles, doesn’t work, and relies heavily on existing projects like Servo and QuickJS. So, ‘from scratch’ apparently means ‘cobbled together from other people’s code by an AI that can’t even make it run.’ Impressive. <a href="https://app.daily.dev/posts/DovPZUpOP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI coding assistants are making you ‘uncreative’: A research paper suggests that AI-powered code generation, or ‘vibe coding,’ is leading to a decline in open-source quality because developers are bypassing maintainers and their feedback. Apparently, the problem isn’t that AI is generating mediocre code, it’s that you are too busy ‘vibe coding’ to give proper feedback. So, if open source dies, it’s your fault for being too productive with AI. <a href="https://app.daily.dev/posts/YRC8HFFhU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI will make your code unmaintainable, but it’s still faster: One developer, after two years of using AI coding agents, found that while the AI generated seemingly high-quality code that passed reviews, it led to ‘cumulative architectural degradation’ and an unmaintainable codebase. The conclusion? Manual coding is ‘ultimately faster and more productive’ when you factor in the full development lifecycle. So, AI helps you write code faster, but then you spend even more time fixing the mess it made. Progress! <a href="https://app.daily.dev/posts/MVxk3C8r1" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>