<h2>Must know</h2>
<ul>
<li><strong>SpaceX acquires xAI and plans orbital AI data centers</strong>: SpaceX has acquired xAI in a deal valued at $250 billion, with plans to build AI data centers in space powered by solar energy. Elon Musk claims this will reduce AI computing costs and leverage Starship rockets for deployment, aiming for 100 gigawatts of AI capacity annually. This consolidation merges Grok chatbot, X platform, Starlink, and SpaceX rockets under one entity, raising concerns about oversight, ethical implications, and national security, especially as xAI is reportedly burning $1 billion monthly. <a href="https://app.daily.dev/feed-by-ids?id=wMYF638eZ&amp;id=n9USqOv7M" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Container security is a mess, and developers are to blame</strong>: A quarter of developers faced container security incidents last year, with human error causing 62% of mistakes. Despite security being a top priority for some, 86% of organizations prioritize performance over security when choosing base images. Many still use bloated general-purpose images, and only a fraction use essential security tools like vulnerability scanning or image signing. This widespread negligence creates a massive attack surface, with Chainguard reporting thousands of CVEs in less popular container images. <a href="https://app.daily.dev/posts/I55IVTrKU" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google disrupts major residential proxy network IPIDEA</strong>: Google’s Threat Intelligence Group successfully disrupted IPIDEA, a massive residential proxy network used by over 550 threat groups for espionage and cybercrime. IPIDEA secretly enrolled millions of consumer devices, including smartphones and Windows PCs, as exit nodes through embedded SDKs. This operation, a collaboration with partners like Spur and Cloudflare, significantly reduced IPIDEA’s device pool and protected Android users via Google Play Protect, highlighting the ongoing battle against sophisticated cyber threats. <a href="https://app.daily.dev/posts/68kBI3xUP" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Deno Deploy reaches GA, launches Sandbox for untrusted code</strong>: Deno Deploy is now generally available, offering zero-config deployment for JavaScript/TypeScript apps with integrated serverless PostgreSQL and OpenTelemetry observability, while Deno Sandbox provides instant, isolated Linux microVMs for securely running untrusted code, especially LLM-generated code. <a href="https://app.daily.dev/posts/Jmdw76tdu" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Snowflake expands with PostgreSQL and Openflow integration</strong>: Snowflake launched a PostgreSQL database-as-a-service for transactional workloads and introduced Openflow, a plug-and-play data integration solution, aiming to unify OLTP and analytics within its AI Data Cloud and reduce ETL complexity. <a href="https://app.daily.dev/feed-by-ids?id=ppuh0CVeq&amp;id=qgdGrzKKG" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are getting a package manager</strong>: Tessl launched a package manager for AI agent skills, aiming to standardize skill management, quality assessment, and version control, treating AI agent capabilities like traditional software packages. <a href="https://app.daily.dev/posts/7qWwPYNkg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI details Codex CLI internals</strong>: OpenAI released insights into its Codex software development agent, detailing its agent loop architecture that constructs prompts from instructions and tools, and manages LLM inference performance through caching and compaction. <a href="https://app.daily.dev/posts/DWHoaTisZ" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>China’s open-source AI ecosystem is booming</strong>: China’s AI ecosystem has shifted to an open-source approach, with major tech companies and startups adopting collaborative development, supported by significant infrastructure investment and an ‘AI+’ action plan prioritizing large-scale deployment over AGI. <a href="https://app.daily.dev/posts/mPvh8h9et" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New terminal UI tool ‘Overitall’ released</strong>: Overitall (oit) is a new terminal UI tool combining process management with advanced log viewing, offering interactive filtering, trace detection, and AI assistant integration for automated log analysis. <a href="https://app.daily.dev/posts/DkhXBDdU9" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel formalizes succession plan for Linus Torvalds</strong>: The Linux kernel project has formalized its first-ever succession plan, outlining a process for maintaining leadership continuity if Linus Torvalds or key maintainers become unavailable, addressing long-term project sustainability concerns. <a href="https://app.daily.dev/posts/vAKZlS9Qz" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>AI coding assistants are secretly transmitting your code to China</strong>: Schneier on Security reports that two popular AI coding assistants, used by 1.5 million developers, are secretly transmitting ingested code to servers in China, allegedly to steal intellectual property. Developers are advised to avoid these tools, because who needs intellectual property when you have ‘AI-powered productivity’? I’m sure the ‘unauthorized data exfiltration’ is just a feature, not a bug. <a href="https://app.daily.dev/posts/puWFcrOO1" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenClaw: the AI agent that’s a security dumpster fire</strong>: OpenClaw, an AI agent project that mysteriously gained 60,000 GitHub stars in 72 hours (definitely not coordinated influencer promotion), has been exposed by security researchers for critical vulnerabilities including data exfiltration and prompt injection. Apparently, bundling existing npm packages with browser session hacking for messaging platforms creates a ‘single point of failure with full system access.’ Who could have seen that coming? Maybe next time, they’ll try actually building something secure instead of a ‘shopping cart full of other people’s work.’ <a href="https://app.daily.dev/posts/fuL6WfKHt" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Microsoft’s AI overload in Windows 11 is being ‘walked back’</strong>: After aggressive AI integration and ‘significant user backlash,’ Microsoft is reportedly scaling back its AI efforts in Windows 11, reviewing Copilot integrations and the infamous Windows Recall. Apparently, users didn’t appreciate their OS becoming a ‘productivity’ nightmare. It’s almost as if forcing half-baked AI features on people without their consent isn’t a winning strategy. Who knew? <a href="https://app.daily.dev/posts/mR53bImT0" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
