<h2>Must know</h2>
<ul>
<li><strong>Apple taps Google Gemini for Siri and AI features</strong>: Apple and Google have inked a multi-year, non-exclusive deal for Apple to integrate Google’s Gemini AI models and cloud infrastructure into its AI features, including Siri. This partnership, reportedly costing Apple around $1 billion annually, comes after Apple evaluated models from OpenAI and Anthropic. The revamped Siri, powered by Gemini, is expected to launch later in 2026, with Apple emphasizing its commitment to privacy despite leveraging Google’s external infrastructure, as reported by Mac-centric and business tech outlets. <a href="https://app.daily.dev/posts/joVK7B34Z" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic launches agentic AI for file management</strong>: Anthropic has released Cowork, a new feature within the Claude Desktop app for macOS, allowing Claude Max subscribers to use agentic AI for file management. This tool enables Claude to autonomously read, edit, and create files within a specified folder via a chat interface, handling tasks like organizing receipts and drafting reports. While it targets non-technical users, Anthropic acknowledges potential risks like prompt injection and accidental file deletion, as detailed in product announcements. <a href="https://app.daily.dev/posts/TLkxIySjV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic restricts third-party access, OpenAI counters</strong>: Anthropic recently restricted OAuth access for third-party coding agents like OpenCode to its Claude Pro/Max subscriptions, forcing them to use higher-cost API access. This move, seen as an attempt to control its value chain, has alienated some users. In response, OpenAI publicly announced official support for using ChatGPT subscriptions with OpenCode and other open-source tools, positioning itself as a more open alternative and intensifying competition in the LLM tool market, according to opinion and technology news sources. <a href="https://app.daily.dev/feed-by-ids?id=8dtXe7P09&amp;id=eRi1VQ5SP" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI data centers drive demand for next-gen nuclear power</strong>: Hyperscale AI data centers, requiring over a gigawatt of electricity each, are fueling a surge of interest in advanced nuclear options like small modular reactors (SMRs) to meet their massive energy demands and reduce carbon emissions. <a href="https://app.daily.dev/feed-by-ids?id=bf0xB32y7&amp;id=qk040xtMi" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google releases interpretability tools for Gemini models</strong>: Google has launched Gemma Scope 2, a suite of interpretability tools for analyzing the internal behavior of Gemini 3 models, using sparse autoencoders and transcoders to understand information processing and identify safety risks. <a href="https://app.daily.dev/posts/uJ4buxaIK" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Postman acquires Fern for API documentation and SDK automation</strong>: Postman has acquired Fern, a startup specializing in automating API documentation and SDK generation in nine programming languages, aiming to reduce developer friction and enhance its end-to-end API lifecycle platform. <a href="https://app.daily.dev/posts/9UqlZV9z7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kubernetes v1.35 introduces workload-aware scheduling</strong>: Kubernetes v1.35 is out, featuring workload-aware scheduling with gang scheduling to prevent partial resource usage and opportunistic batching for faster deployment of identical pods. <a href="https://app.daily.dev/feed-by-ids?id=R6fTuHjeX&amp;id=zTULpLxtX" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft retires MDT, shifts to cloud-dependent deployment</strong>: Microsoft has abruptly retired the Microsoft Deployment Toolkit (MDT) after 20 years, pushing users towards more complex, cloud-integrated solutions like Windows Autopilot or Configuration Manager OSD, leaving many organizations without a transition period. <a href="https://app.daily.dev/feed-by-ids?id=0sSu27x90&amp;id=8WMJpirC0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>MuddyWater APT deploys new Rust-based malware</strong>: The Iran-linked MuddyWater APT group is now using a sophisticated Rust-based malware, RustyWater, in espionage campaigns targeting diplomatic, maritime, financial, and telecom organizations, primarily in the Middle East. <a href="https://app.daily.dev/posts/xFuCzijpN" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux 6.19-rc5 released with driver and security fixes</strong>: Linux 6.19-rc5 is out, bringing fixes for NVIDIA GPUs in the Nouveau driver, addressing a RISC-V security vulnerability, and adding support for the Logitech MX Anywhere 3S mouse. <a href="https://app.daily.dev/posts/MbuLWzK6Y" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI code generation: cheap code, expensive problems</strong>: LLM tools like Claude Code make code generation fast and accessible, even for non-developers. But as Salesforce learned the hard way, replacing 4,000 experienced staff with AI led to reliability issues and a swift reversal. Turns out, building reliable, scalable software still requires actual human expertise, not just cheap, AI-generated lines of code. Who knew? <a href="https://app.daily.dev/feed-by-ids?id=BCKlippws&amp;id=aKgPqgLac" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>India wants smartphone source code, denies it publicly</strong>: India’s proposed smartphone security rules include demands for source code access from manufacturers like Apple and Samsung, along with 12-month device log retention. While government documents reportedly show explicit requirements, the IT ministry publicly denies seeking source code, calling Reuters’ reporting ‘inaccurate.’ So, they want it, but they don’t want you to know they want it. Classic. <a href="https://app.daily.dev/posts/C0BEamxdR" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents: the next big security headache</strong>: AI agents are supposed to be the future, but apparently, they’re just creating new security nightmares. Forget traditional exploits; now we’re dealing with ‘semantic/intent-based attacks’ like GeminiJack and CometJacking. And while 73% of organizations use generative AI, only 4% have a ‘mature’ security posture. So, we’re all rushing into this, but nobody’s actually ready for the inevitable chaos. Good times. <a href="https://app.daily.dev/feed-by-ids?id=CuSSpmwf6&amp;id=EDYtMcTfL&amp;id=JuPUyAfNx" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
