<h2>Must know</h2>
<ul>
<li>Microsoft Office zero-day actively exploited: Microsoft issued an emergency patch for CVE-2026-21509, a zero-day vulnerability in Office with a CVSS score of 7.8, that’s actively being exploited. This flaw bypasses security features for legacy COM and OLE components, allowing attackers to execute code via malicious Office files. While newer Office versions received immediate updates, Office 2016 and 2019 users are stuck with manual registry workarounds for now, as CISA mandates federal agencies patch by mid-February. This is a critical reminder that even widely used software can harbor decades-old vulnerabilities, as reported by multiple sources. <a href="https://app.daily.dev/feed-by-ids?id=qC937DXHe&amp;id=qC937DXHe" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Laravel Livewire RCE vulnerability leads to mass compromises: Numerous Laravel websites have been compromised due to a probable Remote Code Execution (RCE) vulnerability in Livewire versions prior to 3.6.4. Attackers are exploiting improper serialized data handling to execute code, deploying PHP backdoors and CPU cryptominers. The exploit bypasses standard Laravel routing by running files directly from the public/ directory, enabling SEO cloaking and persistent access. Immediate mitigation involves upgrading Livewire and securing application keys, as detailed across several reports. <a href="https://app.daily.dev/feed-by-ids?id=dsQkjC4Yj&amp;id=dsQkjC4Yj&amp;id=dsQkjC4Yj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>ChatGPT caught citing AI-generated encyclopedia: ChatGPT’s GPT-5.2 model has been observed referencing Grokipedia, an AI-generated encyclopedia by Elon Musk, in its responses. The Guardian found nine instances across various topics, raising concerns about recursive AI sourcing amplifying bias and misinformation. OpenAI claims its models filter low-credibility sources, but this incident suggests a significant oversight. This highlights the growing risk of AI models feeding on each other’s output without human verification, potentially leading to ‘model collapse’ and degraded information quality. <a href="https://app.daily.dev/posts/fWRmbZwbT" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li>Ingress-nginx project archived, migrate to Cilium: The ingress-nginx project is being archived in early 2026, ending active maintenance and security updates, pushing users to migrate to Cilium Ingress for a drop-in replacement or Cilium’s Gateway API for advanced features and better performance. <a href="https://app.daily.dev/posts/PptBbhnAu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>WhatsApp adopts Rust for media security: WhatsApp replaced 160,000 lines of C++ with 90,000 lines of Rust in its media handling stack, deploying a new media consistency library called Kaleidoscope to validate media files and detect malicious patterns, enhancing security and performance for billions of users. <a href="https://app.daily.dev/posts/2YIPmtHiF" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Google Cloud launches N4A instances with Axion ARM CPUs: Google Cloud has released N4A compute instances powered by custom Axion ARM processors, offering up to 64 vCPUs and 512GB RAM, targeting web servers, microservices, and containerized applications with improved performance and cost-effectiveness over previous T2A instances. <a href="https://app.daily.dev/feed-by-ids?id=JvBjVyZ5U&amp;id=JvBjVyZ5U" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Open Semantic Interchange specification finalized: The Open Semantic Interchange (OSI) specification v1.0 is now an Apache 2 licensed open standard for sharing vendor-neutral semantic models across data tools, with major players like Databricks, Qlik, and AtScale joining the working group. <a href="https://app.daily.dev/posts/Tmfufk7Hk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Supabase releases AWS PrivateLink database connectivity beta: Supabase now offers PrivateLink in beta for AWS users, enabling private network connections to Supabase databases via AWS VPC Lattice, enhancing security and compliance for regulated industries by eliminating public endpoints. <a href="https://app.daily.dev/feed-by-ids?id=Bgvnt4bF8&amp;id=Bgvnt4bF8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Malicious Chrome extension hijacks Amazon affiliate tags: A Chrome extension named ‘Amazon Ads Blocker’ has been identified as malicious, blocking ads while secretly injecting the developer’s affiliate codes into Amazon product links, diverting commissions from legitimate creators. <a href="https://app.daily.dev/feed-by-ids?id=OavgB4e3i&amp;id=OavgB4e3i" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Vercel CLI adds API command for direct access: Vercel CLI version 50.5.1 introduces a new <code>api</code> command, allowing authenticated HTTP requests to Vercel APIs directly from the terminal, streamlining developer workflows and enabling AI agents like Claude Code to interact with Vercel APIs using inherited user permissions. <a href="https://app.daily.dev/feed-by-ids?id=AvpZcZC0U&amp;id=AvpZcZC0U" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li>AI coding agents: functional but messy: Apparently, AI coding agents are great at churning out code quickly, but they’re not so hot on quality. According to some experiment, they produce ‘spaghetti’ code, ignore conventions, and introduce technical debt, meaning human developers still need to clean up the mess. So much for AI replacing us, looks like we’re just getting more debugging practice. <a href="https://app.daily.dev/feed-by-ids?id=uilUGnhp4&amp;id=uilUGnhp4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>Cursor’s AI browser: a ‘spaghetti’ code masterpiece: Cursor announced a groundbreaking AI-built browser with millions of lines of code and a custom Rust rendering engine. Developers, however, quickly found the code to be ‘spaghetti,’ unable to compile, and heavily reliant on existing projects, completely undermining Cursor’s ‘from scratch’ claims. It seems the AI was great at generating code, just not good code, and Cursor was great at generating hype. <a href="https://app.daily.dev/feed-by-ids?id=gcgtwbk5H&amp;id=gcgtwbk5H&amp;id=gcgtwbk5H" target="_blank" rel="noopener nofollow">Read more</a></li>
<li>AI agents: the new enterprise ‘coworkers’ (with governance issues): AI agents are supposedly moving from experimental tools to embedded ‘coworkers’ in enterprises, shifting from cloud to local environments. This ‘transformation’ brings significant governance challenges, with experts recommending clear autonomy tiers and accountability for agent actions. Because, you know, giving AI more autonomy without proper oversight has never gone wrong before. <a href="https://app.daily.dev/feed-by-ids?id=yBxtGCVqr&amp;id=yBxtGCVqr" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>