<h2>Must know</h2>
<ul>
<li><strong>AI coding tools: productivity gains vs. quality pitfalls</strong>: AI coding assistants like GitHub Copilot and Claude Code show mixed results. While some users report 10-100x productivity gains for well-defined tasks, others, including Ruby on Rails creator David Heinemeier Hansson, find them unreliable and inconsistent, unable to match junior developers in understanding system architecture. Research indicates Copilot boosts velocity by 55% but increases security vulnerabilities by 322% and lowers delivery stability, as reported by multiple sources. The consensus is that AI excels at repetitive tasks but requires significant human oversight and robust quality gates to prevent downstream issues. <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 7003, 8004, 9004, and 9005 series 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. AMD has issued microcode updates, with more scheduled for April 2026, and advises disabling hyperthreading on critical systems as a temporary mitigation. <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 enforcement, which includes new security features like PIN and biometric authentication, has no opt-out and is described as an effective ‘kill switch’ for non-compliant apps, according to multiple 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>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 for multi-modal generative AI tasks. <a href="https://app.daily.dev/posts/XBX3pzO5M" 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 like the Planner Agent and Security Analyst Agent across the software development lifecycle. <a href="https://app.daily.dev/posts/R9WyVAVPv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust gains official Cargo and supply chain support</strong>: Socket announced General Availability for Rust and Cargo, offering dependency analysis, SBOM generation, and supply chain visibility for Rust projects, addressing risks beyond memory safety. <a href="https://app.daily.dev/posts/O8W5FyqYs" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft open-sources XAML Studio</strong>: Microsoft open-sourced XAML Studio under the .NET Foundation, making its lightweight prototyping tool for XAML UI development available on GitHub for community contributions. <a href="https://app.daily.dev/posts/hUJrXK08w" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New language models: GLM-4.7-Flash and Differential Transformer V2</strong>: GLM-4.7-Flash, a 30B-A3B Mixture of Experts model, and Differential Transformer V2, with a revamped attention mechanism, promise efficient deployment and improved training stability for LLMs. <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>Ubisoft sued by former Assassin’s Creed lead</strong>: Marc-Alexis Côté, former Assassin’s Creed lead, is suing Ubisoft for constructive dismissal, claiming he was demoted and forced out after nearly two decades with the company. <a href="https://app.daily.dev/posts/LIV91KSqU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>jQuery 4.0 released after 10 years</strong>: jQuery 4.0, the first major update since 2016, has been released, focusing on removing deprecated features, dropping support for older browsers like IE10, and reducing file size. <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 make you a 100x developer, just ignore the security flaws</strong>: Apparently, AI coding tools like <a href="http://Pythagora.ai" target="_blank" rel="noopener nofollow">Pythagora.ai</a> can generate full-stack apps from prompts in hours, and new books promise to teach beginners game development with AI, no coding required. It’s almost as if the industry wants us to believe that deep technical expertise is now optional. Just don’t ask about the 322% increase in security vulnerabilities or the declining code stability that comes with this ‘progress,’ as some reports conveniently gloss over. <a href="https://app.daily.dev/feed-by-ids?id=LklGkgpse&amp;id=LpF1NSIQw&amp;id=15Mjf12HP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI-powered customer onboarding: because humans are so last year</strong>: Remote is now using LangChain and LangGraph to automate customer onboarding, claiming to reduce data migration times from days to hours. Meanwhile, Clarm, a Y Combinator-backed startup, boasts ‘zero-hallucination AI responses’ for lead generation. It’s truly inspiring how these companies are solving complex problems with AI, as long as you don’t question what ‘zero-hallucination’ actually means or how much human oversight is still required to prevent a complete data apocalypse. <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>Wikipedia needs a ‘Bullshit Police’ for AI-generated content</strong>: Wikipedia has launched ‘WikiProject AI Cleanup’ to combat AI-generated inaccuracies, while a police chief retired after Microsoft Copilot fabricated a report that led to banning Israeli fans from a football match. It seems AI is so good at generating plausible-sounding nonsense that even official institutions need dedicated teams to fact-check it. Perhaps we should just let the AI write all the news, and then have another AI verify it. What could possibly go wrong? <a href="https://app.daily.dev/feed-by-ids?id=1G2Nqa8sb&amp;id=NJTPjA1ik" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
