<h2>Must know</h2>
<ul>
<li><strong>AI is eating open source’s lunch, Tailwind CSS lays off 75% of staff</strong>: Tailwind CSS, despite record adoption, laid off 75% of its engineering team after an 80% revenue drop. The culprit? AI tools, which are now generating Tailwind code and answering developer questions, causing a 40% decline in documentation traffic. This directly impacts Tailwind’s business model, which relies on documentation visits to promote paid products. The maintainer even rejected a pull request to make docs more LLM-friendly, fearing further financial harm. This isn’t just about Tailwind; it’s a stark warning for open-source projects whose monetization depends on content consumption, as reported by multiple sources including The Miners and Hacker News. <a href="https://app.daily.dev/feed-by-ids?id=LLQ1zfubn&amp;id=f3fYMXuKZ&amp;id=RuMhQtj6W&amp;id=tFhv6QoOp&amp;id=d57ILPjBN&amp;id=re8TItYuD&amp;id=IiuW2YECE&amp;id=mpCeVdjI4&amp;id=eugOrAHSX&amp;id=u4S1Vnazf" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft acquires Osmos for AI data engineering in Fabric</strong>: Microsoft just snapped up Osmos, an AI data engineering startup, to automate data preparation within its Fabric platform. This move aims to cut data engineering bottlenecks by using agentic AI to transform raw data into analytics-ready assets. Osmos’s tech will now be exclusive to Fabric, reducing development and maintenance efforts by over 50%, according to Microsoft and analysts. This signals a clear push towards autonomous AI agents for data processing and simplification of data unification. <a href="https://app.daily.dev/posts/dRdzpOrJb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>NVIDIA launches open AI models for autonomous vehicles and local AI acceleration</strong>: NVIDIA unveiled Alpamayo, a family of open-source AI models for autonomous vehicles, including a 10-billion parameter vision-language-action model designed for complex driving scenarios. They also released AlpaSim, an open-source simulation framework. On the local AI front, NVIDIA announced upgrades for RTX-enabled PCs, with ComfyUI, llama.cpp, and Ollama seeing significant inference speed boosts, making local AI development more accessible, as detailed by NVIDIA Developer. This is a big deal for anyone building AI infrastructure or working with autonomous systems. <a href="https://app.daily.dev/posts/UyBvDFk8k" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI coding tools are changing developer workflows</strong>: AI coding assistants like Claude Code are enabling non-programmers to build applications and are shifting developer roles from code producers to supervisors, with Anthropic reporting 59% of developer work now involves AI, leading to 50% productivity gains. <a href="https://app.daily.dev/feed-by-ids?id=FeqXX8Qki&amp;id=QkdkxdhUj&amp;id=QqckxdhUj&amp;id=TaETvLfgQ&amp;id=T7lz8DVGQ&amp;id=ggxtFI8fz&amp;id=zZPpIk0MP&amp;id=mM5up9nt2&amp;id=LeN1TdPNx&amp;id=hHTJdKyX8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel 6.19 and AMD Radeon drivers get major upgrades</strong>: Linux kernel 6.19-rc5 includes a critical fix for the Rust Binder driver deadlock and adds Intel Nova Lake Point S device ID support, while LLVM/Clang 22 now supports Intel Nova Lake, Arm C1, and Ampere1C processors, bringing C2y language features and Zen 4 optimizations. <a href="https://app.daily.dev/feed-by-ids?id=MbuLWzK6Y&amp;id=njv07kgwC&amp;id=DFbxDQnoL" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Rust-based tools for Linux and Python ecosystems</strong>: The Arch Linux Pacman package manager might be replaced by ALPM, a new Rust-based project, while the Python ecosystem sees new Rust-based type checkers like ‘ty’ and packaging tools like ‘uv’ and ‘Hatch’ aiming for better performance and security. <a href="https://app.daily.dev/feed-by-ids?id=w44Nd99pc&amp;id=zSZp3Avnk&amp;id=quf6PiAcD" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kubernetes cost management and IaC tooling get smarter</strong>: OpenCost released 11 versions in 2025, focusing on Kubernetes cost management with Prometheus-optional operation and an AI-powered MCP server for natural language cost queries, while Spacelift Plugins enable native integration of security scanners and cost analyzers into IaC workflows. <a href="https://app.daily.dev/feed-by-ids?id=37oSt4QPm&amp;id=3QpPL9Ryt&amp;id=4bvsJkZO1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents reshape cloud, data, and developer workflows</strong>: Atlassian Rovo now integrates with Databricks AI/BI Genie for conversational data querying, while serverless cloud models are proving incompatible with production-scale AI agents, pushing enterprises towards VPC and on-premises deployments for stateful AI workloads. <a href="https://app.daily.dev/feed-by-ids?id=2XKWuVI4o&amp;id=OcN7tvK3U&amp;id=SXn4AkZZF&amp;id=xM76X8yai&amp;id=xdHmgnzKW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Linux desktop environments and legacy support strategies</strong>: Budgie 10.10 now runs natively on Wayland, dropping X11 support and planning a Qt6-focused Budgie 11, while IceWM 4.0 remains committed to X11, and a new experimental X11 server called Phoenix, written in Zig, emerges to bridge the gap. <a href="https://app.daily.dev/posts/peCpqhbhc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Node.js releases address multiple CVEs</strong>: Node.js released versions 20.20.0 LTS (Iron) and 25.3.0 (Current), addressing multiple CVEs, including disabling futimes with the permission model, adding default TLSSocket/TLS error handlers, and requiring full read/write permissions for symlink API access. <a href="https://app.daily.dev/posts/DwsoMgMKH" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>OpenAI launches ChatGPT Health, because what could go wrong?</strong>: OpenAI just launched ChatGPT Health, a new feature that lets you upload your medical records and wellness app data to get ‘plain-language explanations’ and ‘diet advice.’ They claim it’s encrypted and won’t be used for training, but also explicitly state it doesn’t provide diagnoses and LLMs ‘hallucinate.’ So, you can upload your most sensitive data to an AI that makes things up, but don’t expect it to actually help you when you’re sick. What a time to be alive! <a href="https://app.daily.dev/posts/52NiUSgqi" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents: still not joining the workforce, despite the hype</strong>: Remember all those high-profile tech leaders, like Sam Altman, predicting AI agents would ‘join the workforce’ in 2025? Yeah, that didn’t happen. Apparently, products like ChatGPT Agent underperformed, and experts are still skeptical about current LLM tech being enough for real work automation. But hey, the market for agentic AI is still projected to hit $52 billion by 2030, so keep dreaming those autonomous dreams. <a href="https://app.daily.dev/posts/dYYHmjuRB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Stack Overflow’s question volume plummets, blames AI (and itself)</strong>: Stack Overflow saw a 98% decrease in monthly questions since its 2014 peak, with the decline accelerating right after ChatGPT launched. Developers are saying AI tools are more responsive and less judgmental than Stack Overflow’s ‘toxic culture.’ So, it seems AI is not just taking our jobs, it’s also taking our ability to ask basic questions without getting roasted. Progress! <a href="https://app.daily.dev/feed-by-ids?id=qjJ5kf4y2&amp;id=r1nkPFWDK&amp;id=M7ag65EZ5" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
