<h2>Must know</h2>
<ul>
<li><strong>Microsoft patches actively exploited Windows 0-day</strong>: Microsoft released an urgent patch for CVE-2026-20805, an information disclosure vulnerability in Windows that is already being exploited in the wild. This flaw could allow attackers to leak memory addresses, potentially bypassing ASLR as part of larger attack chains, according to CISA. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) has mandated that federal agencies apply this patch by February 3, underscoring the immediate risk. <a href="https://app.daily.dev/posts/Kwy6WAbQP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents found with critical vulnerabilities in enterprise platforms</strong>: Security researchers uncovered severe vulnerabilities in ServiceNow’s Virtual Agent and Superhuman AI, allowing full system control and email exfiltration, respectively. The flaws stemmed from weak authentication and insecure integration of agentic AI features, enabling attackers to impersonate users or leak sensitive data. While both companies have issued fixes, these incidents highlight significant security risks in AI-driven business productivity tools, as reported by external researchers. <a href="https://app.daily.dev/feed-by-ids?id=2A2nfRTHq&amp;id=EDFS33r2W&amp;id=KcLhdNkuO&amp;id=qah3DxopU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitHub to deprecate older Copilot AI models</strong>: GitHub is deprecating three AI models for Copilot—Claude Opus 4.1, GPT-5, and GPT-5-Codex—on February 17, 2026. Users must update their workflows to integrate the new replacement models (Claude Opus 4.5, GPT-5.2, and GPT-5.1-Codex) to ensure service continuity. This change, communicated directly by GitHub, requires enterprise administrators to manually enable the alternative models in Copilot settings. <a href="https://app.daily.dev/feed-by-ids?id=ImHNcBPhJ&amp;id=ZoKLGVJ9V&amp;id=uw83EGG5r" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Node.js releases security fixes for critical vulnerabilities</strong>: Node.js versions 20.20.0 (LTS) and 25.3.0 (Current) were released with security patches addressing multiple CVEs, including a denial-of-service vulnerability related to stack space exhaustion affecting async_hooks and other frameworks. <a href="https://app.daily.dev/feed-by-ids?id=DwsoMgMKH&amp;id=ImHNcBPhJ&amp;id=ZoKLGVJ9V&amp;id=uw83EGG5r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust in production for embedded smart door locks</strong>: Achilles has been using Rust for 4-5 years in production for smart door lock firmware, leveraging its safety features and efficient code size (300-420KB) with a single firmware engineer, relying on the Embassy async runtime and open-source collaboration. <a href="https://app.daily.dev/feed-by-ids?id=SVl4BYYw0&amp;id=U09VghD1S" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Advanced Linux malware targets cloud and containers</strong>: Check Point Research identified VoidLink, an advanced malware framework written in Zig, designed to target Linux cloud and container environments like Docker and Kubernetes, featuring 37+ modular plugins for credential harvesting and evasion. <a href="https://app.daily.dev/feed-by-ids?id=uJaIfhgHH&amp;id=QWUKVWdsK&amp;id=x9wl564Rc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft responds to AI data center backlash</strong>: Microsoft announced a ‘community-first’ approach to its AI infrastructure expansion, pledging to pay full electricity grid costs, minimize water use, and create local jobs, in response to widespread public opposition across 24 states. <a href="https://app.daily.dev/posts/jqtBl3Jnb" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Confluent acquires DevCycle for feature management</strong>: Dynatrace acquired DevCycle, an OpenFeature-based feature management platform, to integrate progressive delivery methods with its real-time observability tools, aiming to reduce release risk and accelerate cloud-native and AI-native application delivery. <a href="https://app.daily.dev/posts/JSUZtNgZf" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel update brings new developer features</strong>: Laravel 12.47.0 introduces the @includeIsolated Blade directive for variable-isolated template includes, a Cache::withoutOverlapping() method for preventing concurrent cache operations, and enhanced enum support, improving security and maintainability. <a href="https://app.daily.dev/posts/SQ5jDUcaB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New terminal manager for AI coding agents released</strong>: ‘Agent of Empires’ (aoe), a Rust-developed terminal session manager, was released to help developers manage multiple AI coding agent sessions with a TUI dashboard, supporting tools like Claude Code and OpenCode. <a href="https://app.daily.dev/posts/AqojY4mSV" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI is killing open source business models, apparently</strong>: According to some, AI is ‘killing’ open source business models by extracting documentation and reducing web traffic, leading projects like Tailwind CSS to see an 80% revenue decline. So, instead of adapting, open source maintainers are now tightening licenses to ‘limit AI training and corporate exploitation.’ Because nothing says ‘open’ like trying to control how people use your publicly available information. Maybe the problem isn’t AI, but business models that relied on people manually clicking through docs? <a href="https://app.daily.dev/feed-by-ids?id=Ml47Y2Hd6&amp;id=l6Y9rCI4U&amp;id=RRH1cjln0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta’s AI leadership drama: more soap opera than science</strong>: Yann LeCun, Meta’s former chief AI scientist, has reportedly stormed off after Alexander Wang, CEO of Scale AI, was appointed head of superintelligence. LeCun publicly trashed Wang as ‘inexperienced’ and even accused Meta’s AI team of manipulating Llama 4 test results. It sounds less like a strategic pivot and more like a high-school cafeteria fight, but hey, at least it’s entertaining. Who needs actual AI breakthroughs when you have this level of corporate drama? <a href="https://app.daily.dev/feed-by-ids?id=ILAkYeoup&amp;id=PmK6l9z8J" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quantum computing security: because who needs error correction when you have vulnerabilities?</strong>: Penn State researchers are sounding the alarm that cloud-based quantum computers are riddled with security vulnerabilities, from intellectual property theft to undetectable result degradation. They’re urging immediate integration of security, rather than waiting for full error correction. It’s almost as if building incredibly complex, unverified systems and then bolting on security later is a bad idea. Who could have predicted that? <a href="https://app.daily.dev/feed-by-ids?id=JmIuGeLwy&amp;id=MkzC4PEYB&amp;id=QPb7M1IyK&amp;id=dtnIdUSBt&amp;id=f8i188R5R&amp;id=hhNibwkwa&amp;id=yjM5va3Id" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
