<h2>Must know</h2>
<ul>
<li><strong>AI is eating open-source documentation and revenue</strong>: Tailwind CSS reported a 75% staff layoff and an 80% revenue drop, directly blaming AI models for bypassing their documentation site. This means users get answers from AI without visiting the source, undermining Tailwind’s monetization strategy which relies on documentation visits to convert users to paid products. This incident, highlighted by Tailwind CSS’s own statements, exposes a critical challenge for open-source projects that depend on documentation-based revenue, as AI consumes content without contributing to the project’s financial health. <a href="https://app.daily.dev/feed-by-ids?id=RuMhQtj6W&amp;id=f3fYMXuKZ&amp;id=WagqERKHJ&amp;id=UHRxELagh" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding agents are causing bugs and knowledge erosion</strong>: A software consultancy found that while AI tools speed up early ideation, they create review bottlenecks and lead to ‘knowledge erosion’ among developers. AI generates changes faster than humans can review them, and over-reliance on AI output risks developers losing the deep understanding needed for maintenance and evolution. SonarSource also reports that 96% of developers distrust AI-generated code, yet 48% don’t always verify it before committing, leading to security vulnerabilities like SQL injection, especially from models like GPT-5.2 High. <a href="https://app.daily.dev/feed-by-ids?id=DRTczTJyG&amp;id=K8Q8uNU3E&amp;id=Kf8FXuyo&amp;id=E02dc9OJZ&amp;id=q2sjuwRsf" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>North Korean hackers weaponize QR codes for credential theft</strong>: The FBI warns that North Korean state-sponsored hackers (Kimsuky/APT43) are using malicious QR codes in spear-phishing campaigns targeting U.S. think tanks, academic institutions, and government organizations. These ‘quishing’ attacks embed QR codes in emails that redirect users to fake login pages for Microsoft 365, Okta, and VPNs, bypassing traditional email security and multi-factor authentication. This technique exploits the lack of enterprise security on personal mobile devices, making it a significant threat to corporate security practices. <a href="https://app.daily.dev/posts/XmPhFlPf5" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Percona donates database platform to CNCF, Amazon expands Lightsail</strong>: Percona is donating its database management platform, now OpenEverest, to the CNCF for Kubernetes, while Amazon Lightsail introduced larger managed database bundles for production workloads, supporting MySQL and PostgreSQL. <a href="https://app.daily.dev/feed-by-ids?id=MhqrMxmeE&amp;id=waU9VD2zd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New React and Python ecosystem updates</strong>: React 19.2 and the React Compiler are progressing, while Python sees Astral’s ty type checker beta, Django 6, and PythoC for C code generation, alongside a fix for Python 3.14’s tail-calling interpreter optimization. <a href="https://app.daily.dev/feed-by-ids?id=D2DOQ44q3&amp;id=Golang+Weekly+Issue+584%3A+January+9%2C+2026&amp;id=caNcWqICw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>VS Code introduces AI agent skills</strong>: Visual Studio Code now features ‘Agent Skills,’ allowing AI agents like GitHub Copilot to load specialized instructions and resources on demand for domain-specific developer tasks, enhancing automation and context-aware workflows. <a href="https://app.daily.dev/posts/xn90KJxf1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Claude Code 2.1 brings async sub-agents and session teleportation</strong>: Anthropic’s Claude Code 2.1 update introduces forked and asynchronous sub-agents for parallel task execution, automatic skill hot reloading, and ‘session teleportation’ for seamless transitions between terminal and desktop environments. <a href="https://app.daily.dev/feed-by-ids?id=WChBmOscl&amp;id=nDBQDaZMg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Zero Trust principles for AI agents are critical</strong>: Security experts emphasize that Zero Trust principles must treat AI agents as first-class identities with unique credentials and fine-grained authorization, as the Model Context Protocol (MCP) increases security risk by widening context windows. <a href="https://app.daily.dev/feed-by-ids?id=8827svTqM&amp;id=lqqfnNpcw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Flock Safety surveillance network exposes critical vulnerabilities</strong>: A security researcher found hundreds of Flock Safety cameras with publicly exposed admin interfaces and hardcoded ArcGIS API keys, exposing sensitive surveillance data for 12,000 law enforcement and private clients, highlighting systemic credential management failures. <a href="https://app.daily.dev/feed-by-ids?id=3fnWhL9Fd&amp;id=rFxdWrNGg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloud hardware evolution shows uneven progress</strong>: A decade-long analysis of cloud hardware trends reveals significant network bandwidth improvements but modest gains in CPU and memory, with NVMe storage performance stagnating since 2016, driving adoption of disaggregated storage architectures. <a href="https://app.daily.dev/posts/FGn6qNYqW" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Google’s AI-focused Gmail redesign: because you needed more AI in your inbox</strong>: Google is launching an ‘AI-focused Gmail redesign’ with features like AI Overviews to summarize emails and natural language inbox queries. Apparently, your inbox wasn’t proactive enough, so now Gemini 3 will prioritize messages and help you write. Because nothing says ‘productivity’ like an AI deciding what’s important and then writing your emails for you. Just remember, the ‘advanced’ features require a paid subscription. Of course. <a href="https://app.daily.dev/posts/8lOS5k1vH" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cursor AI editor: the VSCode killer that wasn’t</strong>: Cursor, an AI-native code editor, supposedly ‘disrupted’ the coding ecosystem and caused Stack Overflow’s usage to plummet to 2008 levels. While it gained 7 million users, its ‘agent-first development’ and ‘multi-agent systems’ sound impressive until you realize Microsoft restricted its access to the VSCode extension marketplace, and developers are returning to VSCode due to Cursor’s ‘infrastructure instability.’ So much for the revolution. <a href="https://app.daily.dev/feed-by-ids?id=LbaA6unww&amp;id=qjJ5kf4y2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is not replacing developers any time soon (duh)</strong>: Stefan Mischook, a self-proclaimed expert, confidently states that AI is ‘NOT Replacing Developers any time Soon.’ His profound insights include that AI models are ‘pattern recognition systems’ lacking ‘true logical reasoning,’ and they perform poorly on ‘novel problems.’ He also points out that ‘less than 1% bad training data can corrupt results.’ Groundbreaking stuff. It’s almost like we’ve been saying this for years, but thanks for the confirmation, Captain Obvious. <a href="https://app.daily.dev/posts/4FklVf4AA" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
