<h2>Must know</h2>
<ul>
<li><strong>European Parliament bans cloud AI on devices</strong>: The European Parliament has disabled cloud-based AI features on lawmakers’ corporate devices due to data security and privacy concerns, specifically the inability to guarantee data security when AI assistants use cloud services. This temporary ban, affecting tools like writing assistants and summarization, highlights broader European unease about reliance on U.S. tech companies and the handling of sensitive legislative data, despite the EU’s own AI Act. The move prioritizes security and control over convenience, suggesting on-device AI processing as a potential solution. <a href="https://app.daily.dev/posts/NW6nIRqB9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI demand impacts gaming hardware and component shortages</strong>: The insatiable demand from AI datacenters is causing significant memory and storage component shortages, directly impacting the gaming hardware industry. Valve has warned of intermittent stock issues for Steam Deck OLED units and delayed upcoming hardware launches, including the Steam Machine, due to these shortages. Sony may push back the PlayStation 6 until 2028-2029, and Nintendo is considering price increases for the Switch 2, all stemming from the broader RAM crisis fueled by AI. <a href="https://app.daily.dev/posts/ttFo5ZiRW" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI security threats escalate with new exploitation methods</strong>: 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, as demonstrated by Check Point Research. This means AI models can act as runtime decision engines for victim triage and data exfiltration, blending malicious traffic with legitimate enterprise AI usage. Additionally, a novel black-box backdoor attack, LADDER, for convolutional neural networks achieves 99% effectiveness, highlighting the growing sophistication of AI-driven threats. <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" 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, with expanded language support and context windows up to 1 million tokens. <a href="https://app.daily.dev/posts/UOocLbfK4" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OWASP releases smart contract top 10 2026</strong>: OWASP’s Smart Contract Top 10 2026 identifies governance failures, privilege misconfiguration, and upgrade authority concentration as dominant security risks in blockchain systems, emphasizing that audits alone are insufficient for production resilience. <a href="https://app.daily.dev/posts/U7Z7c8wKd" 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>Figma integrates Claude Code for bidirectional design-code workflow</strong>: Figma now allows 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>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>
<li><strong>Meta deploys Nvidia Grace CPUs for AI workloads</strong>: Meta is deploying Nvidia’s Grace CPUs as standalone processors at scale for general-purpose and agentic AI workloads, marking a shift from the typical Grace-Hopper Superchip configuration and delivering 2x performance per watt. <a href="https://app.daily.dev/feed-by-ids?id=tohNH8UoB&amp;id=tohNH8UoB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>DORA compliance becomes commercial imperative</strong>: The Digital Operational Resilience Act (DORA) has transitioned from a regulatory concern to a commercial necessity, influencing supplier due diligence and business deals, requiring organizations to conduct gap assessments and budget for recurring testing cycles. <a href="https://app.daily.dev/feed-by-ids?id=ukJrnq0Ec&amp;id=ukJrnq0Ec" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Gentoo Linux ditches github over ‘copilot-infested’ concerns</strong>: Gentoo Linux is migrating its repository mirrors from GitHub to Codeberg, citing concerns over GitHub Copilot and AI-generated contributions. Apparently, GitHub has ‘struggled with low-quality AI-generated pull requests,’ and some sources even claim GitHub ‘forced Copilot usage.’ Because, you know, a Linux distribution known for its meticulous manual configuration is suddenly worried about ‘quality’ from AI, not just its usual aversion to anything mainstream. <a href="https://app.daily.dev/posts/Ed9ofBB82" 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 ‘trust’ like marketing claims that are definitely not independently verified and sound like they were generated by the very AI they’re selling. <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>Palliser Capital claims Toto 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: toilets are now AI. I guess if you squint hard enough, anything can be an ‘AI play’ if it helps you pump up a stock. Next, they’ll tell us plumbing is the backbone of the AI infrastructure. <a href="https://app.daily.dev/feed-by-ids?id=08TkDSLKo&amp;id=08TkDSLKo" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
