<h2>Must know</h2>
<ul>
<li><strong>Databricks launches Lakebase for AI workloads</strong>: Databricks has introduced Lakebase, a new PostgreSQL-compatible database specifically engineered for AI workloads. This serverless OLTP database integrates seamlessly with Databricks’ data lakehouse architecture, offering features like instant data branching and unified access controls. It’s currently available on AWS, with Azure support in public preview, aiming to bridge operational database capabilities with AI-driven analytics, as reported by Databricks. <a href="https://app.daily.dev/posts/56RHAShlB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google embedding model deprecation causes vendor lock-in</strong>: Google’s deprecation of its <code>text-embedding-004</code> model has rendered thousands of stored vectors unusable for a production RAG-based movie recommender, highlighting a critical vendor lock-in issue. Different embedding models create incompatible vector spaces, making migration difficult and costly. This incident underscores the need to treat embedding models as infrastructure and design for change if persisting embeddings at scale, according to the affected developer. <a href="https://app.daily.dev/posts/xhDI1zj4i" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel 7.0 rc1 released with Rust support and performance gains</strong>: Linus Torvalds has released Linux kernel 7.0 rc1, a version bump attributed to his dislike of large numbers rather than a major feature milestone. This release solidifies Rust support, brings faster cache clearing, non-disruptive kernel updates, and significant performance improvements for IO_uring and memory reclamation. It also includes updated silicon support and RISC-V/LoongArch optimizations, as detailed in various reports. <a href="https://app.daily.dev/feed-by-ids?id=cYP8biu57&amp;id=8ssuMcdSU&amp;id=JXfUDc5F4&amp;id=QdVE1yqfx&amp;id=TcabAvodN&amp;id=cQun6gD6y&amp;id=zaDpYuaxU" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<li><strong>AI coding agents lack innovation, human developers still relevant</strong>: While AI coding agents like Claude Code are highly productive and fast, they fundamentally lack the ability to devise innovative solutions when facing obstacles, a critical differentiator that preserves the relevance of human developers with creative problem-solving skills. <a href="https://app.daily.dev/posts/hB7gw6iXo" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>VS Code extension ‘Fresh File Explorer’ focuses on recent changes</strong>: A new VS Code extension, ‘Fresh File Explorer,’ replaces the standard file explorer to display recently modified files, integrating Git history and offering features like a recency heatmap, ‘Resurrect’ for deleted files, and various grouping modes. <a href="https://app.daily.dev/posts/pQedWHBRB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>OpenAI and Uber reveal infrastructure traffic control platforms</strong>: Uber’s ‘Global Rate Limiter’ handles 80 million requests per second with ‘probabilistic shedding,’ while OpenAI’s ‘Access Engine’ uses a ‘credit waterfall’ to ensure fair access across its systems, both designed for infrastructure-level traffic control. <a href="https://app.daily.dev/posts/OVoWP2oNj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Developer productivity hindered by non-coding tasks</strong>: A developer’s 10-second code fix took over 10 hours due to 14 steps of non-coding tasks, including fighting sandboxed environments, CUDA conflicts, OOM build failures, and compiler segfaults, highlighting how abstraction layers and build systems consume significant time. <a href="https://app.daily.dev/posts/OBIqfWI7x" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Kubernetes 1.35 enhances AI/ML workload support</strong>: Kubernetes v1.35 (‘Timbernetes’) introduces features like workload-aware scheduling with gang scheduling (alpha), stable in-place Pod resource resize, and continued Dynamic Resource Allocation (DRA) progress, solidifying its role as an operating system for AI/ML workloads. <a href="https://app.daily.dev/posts/mAyDHBOpw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Shuru: local-first microVM sandbox for AI agents on macOS</strong>: Shuru is a new macOS tool that uses Apple’s Virtualization.framework to create secure, ephemeral, and reproducible Linux microVMs for AI agent workloads, enabling safe execution of AI-generated code without Docker. <a href="https://app.daily.dev/posts/TqgsBjEoV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust RwLock can be slower than Mutex for read-heavy workloads</strong>: A hands-on investigation found that <code>RwLock</code> performed ~5x worse than <code>Mutex</code> for a read-heavy LRU tensor cache in Rust due to cache-line ping-pong from atomic reader counter increments, suggesting <code>Mutex</code> might be better for very short critical sections. <a href="https://app.daily.dev/posts/rFMs5918e" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic’s Claude compiler video: more marketing than reality</strong>: Anthropic released a marketing video showcasing Claude autonomously building a C compiler capable of compiling Linux and Doom. Critics are calling it ‘dishonest framing,’ suggesting the real goal was a stress test of Claude’s autonomous runtime, not a straightforward compiler-building feat. So, it seems Claude is great at compiling hype, if nothing else. <a href="https://app.daily.dev/posts/vGYk0DbTO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta patents AI to simulate deceased users: because we needed more digital ghosts</strong>: Meta was granted a patent for an AI system designed to simulate deceased users’ social media activity, including posting and chatting, using an LLM trained on their past data. Because nothing says ‘grief counseling’ like a chatbot pretending to be your dead relative. Ethical and mental health concerns? Nah, just more ‘engagement’ for the metaverse. <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Neiry’s ‘bio-drones’: pigeons with neural interfaces for surveillance</strong>: Neiry is testing ‘bio-drones’—live pigeons fitted with neural interfaces, solar panels, and cameras—to guide them along preset routes for infrastructure inspection and search-and-rescue. They claim ‘privacy compliance’ through on-device filtering, but let’s be real, it’s pigeons with brain implants and cameras. What could possibly go wrong with a flock of government-controlled, AI-guided spy birds? <a href="https://app.daily.dev/posts/m7v8MA0Dh" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
