<h2>Must know</h2>
<ul>
<li><strong>Apple and Google team up for AI, Anthropic restricts access</strong>: Apple is partnering with Google to integrate Gemini models and cloud infrastructure for a major Siri overhaul, with Apple reportedly paying Google $1 billion annually. This non-exclusive deal expands their existing relationship, sidelining OpenAI’s ChatGPT for complex queries, as reported by TechCrunch. Meanwhile, Anthropic has restricted third-party CLI tools like OpenCode from accessing Claude Pro/Max subscriptions, forcing developers to use their pay-per-use API. OpenAI, in contrast, is supporting ChatGPT subscriptions within OpenCode, intensifying the rivalry for developer ecosystems. <a href="https://app.daily.dev/feed-by-ids?id=joVK7B34Z&amp;id=8dtXe7P09&amp;id=eRi1VQ5SP" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI data centers demand massive power, nuclear energy steps up</strong>: Hyperscale AI data centers, housing hundreds of thousands of GPUs, are consuming over 1 gigawatt of electricity each, equivalent to a small city, with over 50% of this power still from fossil fuels. This immense demand is driving innovation in next-generation nuclear reactors, including small modular reactors (SMRs) and advanced fuels, designed for safer, quicker deployment to meet the growing energy needs of AI and economic growth. Meta, for instance, has already secured 6.6 GW of nuclear energy by 2035 to power its expanding AI infrastructure, as highlighted by MIT Technology Review. <a href="https://app.daily.dev/feed-by-ids?id=bf0xB32y7&amp;id=qk040xtMi&amp;id=dJDy4tUxK&amp;id=oJJhQtHHE" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New Linux malware in Zig targets cloud, Rust gains ground</strong>: Check Point Research has uncovered VoidLink, a sophisticated Linux malware framework written in Zig, designed to target cloud environments by detecting major providers like AWS and Kubernetes. This framework includes rootkit capabilities and adaptive stealth, indicating a growing threat to cloud infrastructure. Concurrently, the Iran-linked APT group MuddyWater is deploying new Rust-based malware, ‘RustyWater,’ in espionage campaigns, replacing older PowerShell tools. These developments underscore the increasing use of modern, memory-safe languages in advanced cyber threats, while Rust continues to see adoption in critical systems like the Linux kernel. <a href="https://app.daily.dev/feed-by-ids?id=xFuCzijpN&amp;id=iFUbypHN8&amp;id=iaM6CYeL0" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>Oracle’s Java roadmap focuses on performance and concurrency</strong>: Oracle’s 2026 Java roadmap includes Project Valhalla for value types, Project Leyden for AOT compilation, and Project Loom for structured concurrency, aiming to boost speed, scalability, and interoperability for machine learning workloads. <a href="https://app.daily.dev/posts/gmrJODR16" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux desktop environments diverge on display servers</strong>: IceWM 4.0 continues active development on X11 with new features, while Budgie 10.10 fully transitions to Wayland-native operation, and Budgie 11 plans a rewrite using Qt6 and KDE Frameworks, showcasing a fragmentation in the Linux GUI space. <a href="https://app.daily.dev/posts/peCpqhbhc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Postman acquires Fern for API documentation and SDK automation</strong>: Postman acquired Fern, a startup specializing in automated API documentation and SDK generation across nine languages, aiming to reduce developer friction and expand Postman’s full API lifecycle offerings. <a href="https://app.daily.dev/posts/9UqlZV9z7" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Neuron AI Laravel SDK brings agentic AI to PHP</strong>: The Neuron AI Laravel SDK has been released, enabling PHP developers to create autonomous AI agents within the Laravel framework, supporting multiple AI providers like Anthropic, OpenAI, and Gemini without requiring Python. <a href="https://app.daily.dev/feed-by-ids?id=4bt4Q0nem&amp;id=QVuzuBqAB&amp;id=pfLq9n48m" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>GoBruteforcer botnet targets crypto and Linux servers</strong>: The GoBruteforcer malware is exploiting weak credentials, including those from LLM-generated tutorials, to target cryptocurrency projects like TRON and Linux servers via FTP, MySQL, PostgreSQL, and phpMyAdmin services. <a href="https://app.daily.dev/posts/nO4oxH8nA" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Bun v1.3.6 boosts performance and adds new APIs</strong>: Bun v1.3.6 introduces new APIs for tarball handling and JSONC parsing, significantly improving performance with 3.5x faster Response.json() and 15% faster async/await, alongside numerous bug fixes and Node.js compatibility enhancements. <a href="https://app.daily.dev/posts/Wx4br4Zyw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Auto-CPUFreq 3.0 enhances Linux battery and CPU control</strong>: Auto-CPUFreq 3.0 for Linux now offers CPU turbo mode override via CLI and GUI, improved battery device configuration, and specific support for ASUS laptop charge thresholds, enhancing battery life and system performance management. <a href="https://app.daily.dev/posts/umWMqgJxW" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Google’s AI Overviews: dangerous health advice, but mostly accurate, trust us</strong>: Google’s AI Overviews were caught giving dangerous and inaccurate medical advice, specifically on liver test results, which could mislead patients into skipping critical care. Google’s response? They disabled <em>some</em> problematic searches, but many others with similar risks are still out there. They claim most AI Overviews are accurate and clinician-reviewed, but apparently, their clinicians missed the part where bad advice could kill someone. Nothing to see here, folks, just move along. <a href="https://app.daily.dev/posts/fofUhg4zp" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s Cowork: agentic AI for non-coders, what could go wrong?</strong>: Anthropic launched Cowork, an ‘agentic AI tool’ for non-coders, allowing Claude to autonomously read, edit, and create files in a selected folder. It’s in ‘research preview’ and exclusive to macOS Claude Max subscribers. They mention risks like prompt injection and unintended file deletion, but don’t worry, ‘user control is maintained.’ Because giving an AI autonomous file system access with a chat interface to ‘organize receipts’ sounds like a perfectly safe idea for non-technical users. <a href="https://app.daily.dev/posts/TLkxIySjV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI coding tools: code is cheap, but software is still expensive, apparently</strong>: LLM tools like Claude Code have made generating code ‘trivial,’ but building robust, maintainable software remains ‘expensive and complex.’ The new narrative is that AI can’t replace ‘judgment, domain expertise, and technical oversight.’ So, after all the hype about AI writing all the code, it turns out we still need engineers for the ‘hard’ stuff. Who would’ve thought? It’s almost like the value was never in typing characters, but in understanding what to type. <a href="https://app.daily.dev/posts/BCKlippws" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
