<h2>Must know</h2>
<ul>
<li><strong>AI coding tools: hype vs reality</strong>: Despite the hype, AI coding tools like GitHub Copilot are showing mixed results. While they can boost velocity, they also lead to lower delivery stability and significantly more security vulnerabilities, according to research cited in multiple articles. Ruby on Rails creator David Heinemeier Hansson and AWS CEO Matt Garman both express concerns that AI lacks the architectural understanding of even junior developers, making its output unreliable for complex systems. This suggests that while AI can handle routine tasks, deep human expertise in software fundamentals remains critical, as highlighted by Phil Eaton. <a href="https://app.daily.dev/feed-by-ids?id=0fzwUL3Um&amp;id=15Mjf12HP&amp;id=2IKuR9DJZ&amp;id=3DI7vnha4&amp;id=3sM5oeueQ&amp;id=AJzG7x8FZ&amp;id=AS53pzdrZ&amp;id=ATZPMmQn8&amp;id=DdiXGTqPN&amp;id=EqMZfdLoq" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AMD Zen processors vulnerable to StackWarp attack</strong>: A new hardware vulnerability, StackWarp (CVE-2025-29943), affects AMD Zen generations 1 through 5, including EPYC processors. This flaw allows privileged attackers to execute arbitrary code within SEV-SNP protected confidential virtual machines by exploiting an undocumented control bit in the CPU’s stack engine. While AMD has issued microcode updates, additional patches are expected in April 2026, and disabling hyperthreading is advised for critical systems, as reported in the technical vulnerability disclosure. <a href="https://app.daily.dev/posts/9DURDnZX3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft Intune enforces mandatory app updates</strong>: Microsoft Intune will require administrators to update iOS and Android enterprise apps to the latest SDK versions by January 19, 2026. Apps failing to meet these requirements will be blocked from launching, impacting both Microsoft and third-party applications. This move, which includes new security features like PIN enforcement and biometric authentication, has no opt-out and acts as an effective ‘kill switch’ for non-compliant apps, according to reports. <a href="https://app.daily.dev/posts/PNEoUht10" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Rust expands platform support and redefines value</strong>: Socket now offers dependency analysis and SBOM generation for Rust projects, addressing supply chain threats beyond memory safety, while a keynote at NDC TechTown 2025 argues Rust’s primary value lies in program correctness enforced by its static type system, not just memory safety. <a href="https://app.daily.dev/feed-by-ids?id=O8W5FyqYs&amp;id=UzbpNQhI8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GitLab launches Duo Agent Platform for AI orchestration</strong>: GitLab 18.8 introduces the Duo Agent Platform, enabling orchestration of multiple AI agents across the software development lifecycle, including a Planner Agent for project tasks and a Security Analyst Agent for automated vulnerability triage. <a href="https://app.daily.dev/posts/R9WyVAVPv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft open-sources XAML Studio</strong>: Microsoft has open-sourced XAML Studio under the .NET Foundation, making its lightweight prototyping tool for XAML UI development available on GitHub for community contributions, with version 2.0 featuring a UI redesign and deeper WinUI 3 integration. <a href="https://app.daily.dev/posts/hUJrXK08w" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Intel updates LLM-Scaler-Omni for Arc Graphics</strong>: Intel released an update to its LLM-Scaler-Omni product, enhancing support for ComfyUI and SGLang on Intel Arc Graphics hardware, targeting multi-modal generative AI capabilities for image, voice, and video generation. <a href="https://app.daily.dev/posts/XBX3pzO5M" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft patches critical Windows 11 shutdown bug</strong>: Microsoft issued an emergency out-of-band update (KB5077797) for Windows 11 23H2 to fix a critical bug from the January Patch Tuesday that prevented proper shutdown, restart, or hibernation, also addressing Remote Desktop authentication issues. <a href="https://app.daily.dev/feed-by-ids?id=u2UYa6XNM&amp;id=KR0gCuX2r" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New language models GLM-4.7-Flash and Differential Transformer V2 emerge</strong>: GLM-4.7-Flash, a 30B-A3B Mixture of Experts (MoE) model, aims for efficient deployment and claims superior benchmark performance, while Differential Transformer V2 introduces a revamped attention mechanism for improved training stability and efficiency. <a href="https://app.daily.dev/feed-by-ids?id=0sHlXeafO&amp;id=a35LKDHC2" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>jQuery 4.0 released after 10 years</strong>: jQuery 4.0 has been released, marking its 20th anniversary with a major cleanup that removes deprecated features, drops support for Internet Explorer 10 and older, and rewrites the library into ES modules, focusing on modernization rather than new APIs. <a href="https://app.daily.dev/feed-by-ids?id=5SOCKxGlX&amp;id=dThF1vPz2" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI will solve all your customer onboarding woes, apparently</strong>: Remote is touting its use of LangChain and LangGraph to automate customer onboarding, claiming it reduces data migration times from days to hours and bypasses LLM hallucination risks by using Python in a WebAssembly sandbox. Because nothing says ‘cutting-edge AI’ like having an LLM decide what transformations to perform, then letting good old Python do the actual work. It’s almost as if the ‘AI’ part is just a fancy wrapper for well-engineered data pipelines. <a href="https://app.daily.dev/feed-by-ids?id=LVS7zKWbX&amp;id=x5GAIQ3mP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-ready JavaScript formatting: because your AI needs a linter too</strong>: <a href="http://Ultracite.ai" target="_blank" rel="noopener nofollow">Ultracite.ai</a> is pushing a ‘zero-config, AI-ready’ formatter and linter for JavaScript/TypeScript, built on Rust, that integrates with AI coding assistants to ensure generated code matches project standards. Apparently, AI-generated code is so consistently bad that we now need specialized tools to make sure it adheres to our style guides. It’s a bold new world where we’re not just debugging human code, but also the AI’s interpretation of our preferences. <a href="https://app.daily.dev/posts/aqdtePQaS" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-powered game development: just add prompts and stir</strong>: A new book promises to teach ‘vibe coding’ for video games using Python and AI assistants like Gemini and Copilot, claiming even complete beginners can build classic game clones with just AI prompts. Because who needs to learn actual programming fundamentals when you can just ask an LLM to do it for you? This sounds like a fantastic way to churn out unmaintainable, black-box code that will inevitably break the moment you try to do anything slightly complex. <a href="https://app.daily.dev/feed-by-ids?id=LklGkgpse&amp;id=LpF1NSIQw" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
