<h2>Must know</h2>
<ul>
<li><strong>APT28 exploits Microsoft Office zero-day</strong>: The Russia-linked APT28 group is actively exploiting a newly disclosed Microsoft Office zero-day vulnerability (CVE-2026-21509), targeting Ukrainian government agencies and EU organizations. Attacks began within days of Microsoft’s disclosure, using weaponized documents to deploy the COVENANT framework and route traffic through legitimate cloud storage to evade detection. CERT-UA warns that slow update cycles will likely lead to increased attacks, emphasizing the urgency of applying available patches. <a href="https://app.daily.dev/posts/7E37fbDrv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Notepad++ update infrastructure compromised</strong>: Notepad++'s update infrastructure was compromised by suspected Chinese state-sponsored hackers from June to December 2025, delivering backdoored versions to select targets. Attackers exploited a vulnerability in the shared hosting server, intercepting update traffic and maintaining credential access even after the vulnerability was patched. Notepad++ has since released hardened versions (8.8.8.8, 8.8.9, 8.9) with stronger signature verification and migrated to a new hosting provider, urging users to update and remove previously installed self-signed root certificates. <a href="https://app.daily.dev/posts/diIJMavHg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Windows emergency patch frequency increases</strong>: Microsoft released two out-of-band emergency patches in early 2026, continuing a trend where nearly every Patch Tuesday requires subsequent fixes. This increasing frequency of ‘atypical’ emergency releases, coupled with Microsoft’s workforce reductions and claims of 30% AI-contributed code, raises concerns about quality control. System administrators face the dilemma of delaying critical security updates or risking productivity losses from broken patches, with Microsoft remaining silent on the issue. <a href="https://app.daily.dev/posts/H1aWwdchu" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>OpenAI launches macOS app for Codex AI agent</strong>: OpenAI released a macOS desktop app for Codex, its AI coding agent, shifting focus to managing multiple parallel AI agents for complex, long-running tasks, integrating with GitHub and supporting Git worktrees. <a href="https://app.daily.dev/posts/c2BjVZb6r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Firefox 148 introduces granular AI controls</strong>: Firefox 148, releasing in late February, will include a new ‘AI Controls’ panel allowing users to disable generative AI features at a granular level or entirely, addressing previous user backlash and emphasizing optionality. <a href="https://app.daily.dev/posts/K3PNUlsQj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding tools impact software development</strong>: AI coding assistants increased code production speed by 55% but also raised code churn and review times, shifting the bottleneck from writing to verification and necessitating deterministic guardrails and platform engineering for quality. <a href="https://app.daily.dev/posts/KYToiOjQs" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New developer tools for AI integration and UI development</strong>: New tools like <a href="http://ui.sh" target="_blank" rel="noopener nofollow">ui.sh</a> for AI-assisted UI development, Vibe for lightweight Linux VMs on ARM Macs for LLM agents, and Reposit.bot for sharing AI conversation solutions are emerging to streamline developer workflows. <a href="https://app.daily.dev/feed-by-ids?id=FBZjjQkX5&amp;id=XJySsKWNc&amp;id=YD299xQxI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux From Scratch drops SysVinit support</strong>: Linux From Scratch (LFS) will become systemd-only starting with version 13.0 due to overwhelming volunteer workload and increasing upstream dependencies on systemd from major desktop environments. <a href="https://app.daily.dev/posts/ohWaYbFxO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Git 2.53 default-enables Rust support</strong>: Git 2.53 was released with performance optimizations and bug fixes, notably default-enabling Rust support in its build systems, with Rust becoming mandatory in Git 3.0 by late 2026. <a href="https://app.daily.dev/posts/st4SfbY5r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Arcjet JS SDK v1.0 released</strong>: Arcjet has released v1.0 of its JavaScript SDK, emphasizing stability, predictable monthly releases, and minimal breaking changes to provide a production-ready API for security tooling without developer fatigue. <a href="https://app.daily.dev/posts/a40JTbiKw" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI-related layoffs are just ‘AI-washing’</strong>: A Forrester report claims over 50,000 layoffs in 2025 were attributed to AI, but many companies are just ‘AI-washing’—blaming AI for job cuts actually driven by financial pressures or pandemic over-hiring. So, AI is the new scapegoat for bad business decisions? Convenient. Investors love the sound of ‘AI efficiency,’ even if it’s just a fancy term for ‘we messed up our hiring projections.’ <a href="https://app.daily.dev/posts/yXHHUa4yo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Cloudflare’s ‘production-grade’ matrix homeserver</strong>: Cloudflare published a blog post claiming a ‘production-grade’ Matrix homeserver built on Workers, which Matrix co-founder Matthew Hodgson quickly debunked as lacking critical features like federation, proper authentication, and encryption. Apparently, ‘production-grade’ now means ‘barely functional proof-of-concept with unreviewed AI-generated code.’ Good to know the bar is so low. <a href="https://app.daily.dev/posts/Iq2RG6GgT" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated content persists despite exposure</strong>: Research suggests AI-generated and manipulated content continues to influence beliefs even after being exposed as fake, with tools like Adobe’s Content Authenticity Initiative failing to prevent manipulation. So, we’re saying people are too dumb to un-believe something once they’ve seen it, even when told it’s a lie? Sounds like a solid foundation for our AI-powered future. What could possibly go wrong? <a href="https://app.daily.dev/posts/savTmPDZC" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
