<h2>Must know</h2>
<ul>
<li><strong>Critical Node.js DoS vulnerability impacts major frameworks</strong>: A critical denial-of-service vulnerability (CVE-2025-59466) in Node.js, affecting versions 8.x through 24.x, can cause process crashes if <code>async_hooks</code> or <code>AsyncLocalStorage</code> are enabled. This impacts major frameworks like Next.js and React Server Components, as well as APM tools such as Datadog, New Relic, and OpenTelemetry. Node.js has released patches for supported versions, and immediate upgrades are recommended to prevent server termination and service downtime, as confirmed by Node.js release notes and Datadog advisories. <a href="https://app.daily.dev/posts/EyLXGRUe2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ServiceNow AI agent vulnerability allowed full user impersonation</strong>: A critical vulnerability, dubbed ‘BodySnatcher’ (CVE-2025-12420), was discovered in ServiceNow’s Virtual Agent AI platform. This flaw allowed unauthenticated attackers to impersonate any user with just an email address, bypassing passwords and MFA. The exploit could lead to admin account creation, access to connected ServiceNow modules, and privilege escalation across integrations. ServiceNow has since issued a patch, rotated credentials, and removed the exploited AI agent functionality, as reported by AppOmni and ServiceNow. <a href="https://app.daily.dev/posts/wzAqd15ET" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Patch Tuesday addresses critical zero-day and Secure Boot flaws</strong>: Microsoft’s January 2026 Patch Tuesday includes critical security updates for 112-114 vulnerabilities. Among them is CVE-2026-20805, a zero-day in Desktop Window Manager actively exploited to bypass ASLR, which CISA has added to its Known Exploited Vulnerabilities catalog. The update also fixes certificate expiration issues in Secure Boot from 2011 and addresses three certificates expiring in 2026, preventing potential boot process compromises. This is a big one, so get patching, as Krebs on Security and BleepingComputer advise. <a href="https://app.daily.dev/posts/Kwy6WAbQP" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Cursor AI editor optimizes token usage with dynamic context discovery</strong>: Cursor, an AI code editor, introduced dynamic context discovery to reduce LLM token consumption by 46.9% by only accessing code context when needed, aiming for faster AI responses and lower operational costs. <a href="https://app.daily.dev/feed-by-ids?id=gz0G04ehE&amp;id=nGGCLLBtk" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Firefox 147 and Thunderbird 147 released with new features</strong>: Firefox 147 adds automatic Picture-in-Picture, reorganized tab settings, enhanced WebGPU for Apple Silicon, and zero-copy hardware video playback for AMD GPUs, while Thunderbird 147 introduces a ‘Show Full Path’ option and optional folder-name localization. <a href="https://app.daily.dev/posts/f7pzHnDhN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Databricks partners for AI-driven data insights and retail planning</strong>: Databricks is expanding its AI capabilities through partnerships with Atlassian Rovo for natural language queries on enterprise data and Algonomy for AI-driven retail planning solutions, aiming to democratize data access and optimize retail operations. <a href="https://app.daily.dev/feed-by-ids?id=2XKWuVI4o&amp;id=8pU7lXz1A" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Laravel Boost v1.8.10 adds AI prompts and Livewire v4 support</strong>: Laravel Boost v1.8.10 introduces new AI-powered prompts for code simplification and Livewire v4 upgrades, inspired by the Claude plugin, alongside improved installation documentation. <a href="https://app.daily.dev/posts/TQr7IJii0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rails engine launched for LLM API monitoring</strong>: RubyLLM::Monitoring is a new Rails engine that tracks LLM API usage, cost, performance, and error rates within Rails applications, offering a dashboard and configurable alerts for cost management and optimization. <a href="https://app.daily.dev/feed-by-ids?id=oYtsCmpWz&amp;id=ypxl4mcW2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Intel Compute Runtime adds Crescent Island and Nova Lake S GPU support</strong>: Intel’s Compute Runtime version 26.01.36711.4 introduces initial support for Crescent Island and Nova Lake S graphics hardware, continuing to support Level Zero and OpenCL APIs. <a href="https://app.daily.dev/posts/54Uqeadiv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google Trends and Kaggle enhance AI usability and evaluation</strong>: Google Trends now integrates Gemini AI for identifying related search trends and supports up to eight simultaneous terms, while Kaggle launched Community Benchmarks for custom, community-driven AI model evaluation tasks. <a href="https://app.daily.dev/feed-by-ids?id=WhUrXR4tI&amp;id=XLEr4o2U6" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI agents: still not senior engineers, but great for debugging your Sentry integration</strong>: Apparently, Claude Opus 4.5 is fantastic at executing tasks within ‘well-architected systems’ like debugging Sentry or migrating AWS ECS. But don’t expect it to replace your senior engineers who actually design those systems, because it ‘lacks senior-level judgment and abstraction skills.’ So, it’s great for the grunt work, but still needs a human to tell it what to do. Shocking, I know. <a href="https://app.daily.dev/feed-by-ids?id=GMZHoSF60&amp;id=Jj0H2b0XQ&amp;id=SMEJx6ENk&amp;id=Z1R08OC6I&amp;id=ZBZXWSG4p&amp;id=ZsPWeN0Lt&amp;id=dl9u8nkjK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-generated books: the new frontier of technical ‘knowledge’</strong>: HiTeX Press has managed to churn out over 800 ‘technical books’ in a single year, all ‘entirely by AI without human oversight.’ These masterpieces feature ‘fabricated code examples’ and ‘non-existent programming features,’ making them perfect for anyone looking to learn absolutely nothing. And the best part? They’re available on major platforms like Amazon, polluting the technical publishing landscape with industrial-level spam. Who needs quality control when you have quantity, right? <a href="https://app.daily.dev/posts/PmK6l9z8J" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Bandcamp bans AI music, because apparently, humans still make better noise</strong>: Bandcamp, bless their hearts, has decided to ban ‘wholly or substantially AI-generated music’ to ‘protect human musicians’ and ensure ‘music authenticity.’ This, of course, comes after AI-generated songs are already hitting major charts and securing record deals, and companies like Suno are raising hundreds of millions despite copyright lawsuits. It’s almost like Bandcamp realized their revenue model relies on actual human connection, unlike those other platforms drowning in AI-generated noise. Good for them for trying to keep it real, I guess. <a href="https://app.daily.dev/feed-by-ids?id=LZT0kWKSj&amp;id=wgoMt7tqL" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
