<h2>Must know</h2>
<ul>
<li><strong>European Parliament bans cloud AI on devices</strong>: The European Parliament has disabled AI features on lawmakers’ corporate devices due to significant data security and privacy concerns. Their IT department can’t guarantee data security when AI assistants use cloud services, fearing confidential information could be processed in unknown locations, especially given many AI company servers are US-based. This move highlights a pragmatic approach to AI adoption, prioritizing security and control over convenience, even as the EU regulates AI via its comprehensive AI Act. <a href="https://app.daily.dev/posts/NW6nIRqB9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI is accelerating software development, but also technical debt</strong>: AI coding tools are making it easier to generate code, but this speed amplifies existing problems if engineering fundamentals like testing, security, and CI/CD are weak, turning AI into a ‘technical debt accelerator.’ Senior engineers are now tasked with mentoring newcomers who rely heavily on AI, ensuring architectural rigor and guardrails. Research also shows that overly detailed context files for AI agents can actually reduce success rates and increase inference costs, suggesting that minimal, essential requirements are more effective. <a href="https://app.daily.dev/feed-by-ids?id=S6u98Bbpq&amp;id=ZrBLHCGzU&amp;id=vtqAmTWpP&amp;id=xGLYUerTw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI security threats are escalating with new attack vectors</strong>: Check Point Research demonstrated how AI assistants like Grok and Microsoft Copilot can be exploited as covert Command-and-Control (C2) proxies without API keys, enabling attackers to tunnel malware communications. This blends malicious traffic with legitimate enterprise AI usage, making detection harder. Additionally, a critical RCE vulnerability was found in a popular WordPress plugin, allowing unauthenticated attackers to upload and execute arbitrary PHP files, highlighting the ongoing risks in widely used software components. <a href="https://app.daily.dev/feed-by-ids?id=8kxTLeIWY&amp;id=DCxMGa8ig&amp;id=H92K14NPx&amp;id=JY689C6zi&amp;id=K2suxKLLr&amp;id=T8PHWMs0E&amp;id=duMdZgpLR&amp;id=fq2jK89o7&amp;id=ncHmHkLQS&amp;id=rObMrNtnR" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Alibaba releases Qwen3.5 multimodal AI model</strong>: Alibaba launched Qwen3.5, a new multimodal AI model designed for enterprise agent workflows, claiming benchmark improvements over GPT-5.2, Claude 4.5 Opus, and Gemini 3 Pro, and expanding language support to 201 languages. <a href="https://app.daily.dev/posts/UOocLbfK4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Moonshot AI launches Kimi K2.5 multimodal LLM</strong>: Moonshot AI released Kimi K2.5, an open-weight multimodal LLM with vision capabilities and an Agent Swarm mode that can orchestrate up to 100 sub-agents, claiming superior performance on research benchmarks. <a href="https://app.daily.dev/posts/tfdJsRaBj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Figma integrates Claude Code for bidirectional design-code workflow</strong>: Figma now supports converting live code interfaces from Claude Code into editable Figma designs, enabling a bidirectional workflow from code-first exploration to collaborative canvas-based design iteration. <a href="https://app.daily.dev/feed-by-ids?id=iBUqWA8OS&amp;id=jwuzwzVmy" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gentoo Linux migrates from GitHub due to AI concerns</strong>: Gentoo Linux is moving its repository mirrors from GitHub to Codeberg, citing concerns over GitHub Copilot and AI-generated contributions, maintaining a strict policy against AI-assisted code due to copyright and quality issues. <a href="https://app.daily.dev/posts/Ed9ofBB82" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Temporal secures $300M for AI and workflow orchestration</strong>: Temporal raised $300M at a $5B valuation, driven by demand for Durable Execution in AI and long-running workflows, with companies like OpenAI using it for AI agent orchestration. <a href="https://app.daily.dev/feed-by-ids?id=Btt27SjiL&amp;id=H860fyfMj&amp;id=HqHJFpyjl&amp;id=ME7WSXyrj&amp;id=O7fDdRA8l&amp;id=X1zVEwts3" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OWASP releases Smart Contract Top 10 2026</strong>: OWASP released its Smart Contract Top 10 2026, highlighting governance failures, privilege misconfiguration, and upgrade authority concentration as dominant security risks in blockchain systems, based on hundreds of millions in losses from 2025 exploits. <a href="https://app.daily.dev/posts/U7Z7c8wKd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 7.0 optimizes close_range system call</strong>: Linux 7.0 introduces a performance optimization for the <code>close_range</code> system call, reducing its complexity from O(range size) to O(active file descriptors) for significantly faster operations on sparse file descriptor tables. <a href="https://app.daily.dev/feed-by-ids?id=2IrhD5aAX&amp;id=2IrhD5aAX" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI is causing component shortages for gaming hardware</strong>: Valve is blaming ‘AI datacenter demand’ for ‘ongoing memory (RAM) and storage component shortages’ impacting Steam Deck OLED units and delaying other hardware. So, apparently, all those AI models are eating up all the RAM, leaving us gamers in the lurch. Because, you know, it’s not like there are other factors in global supply chains, it’s always AI’s fault now. <a href="https://app.daily.dev/posts/ttFo5ZiRW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quest Software launches AI-powered data management platform</strong>: Quest Software launched its ‘Trusted Data Management Platform’ with an ‘Automated Data Product Factory’ that uses generative AI to create data products from natural language prompts. They claim it delivers ‘up to 54% faster time to data product delivery’ and ‘40% cost savings.’ Because nothing says ‘trusted’ like marketing claims that are definitely not independently verified. <a href="https://app.daily.dev/feed-by-ids?id=wlE5eAAsI&amp;id=wlE5eAAsI" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Toilet maker is an ‘AI play’</strong>: Activist investor Palliser Capital is trying to convince everyone that Japanese toilet manufacturer Toto is an ‘overlooked AI play’ because its advanced ceramics division makes electrostatic chucks for 3D NAND memory chips. Yes, you read that right. A toilet company is now an AI company. Because in 2026, everything is an AI company if you squint hard enough and have a financial incentive. <a href="https://app.daily.dev/feed-by-ids?id=08TkDSLKo&amp;id=08TkDSLKo" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
