<h2>Must know</h2>
<ul>
<li><strong>Databricks launches lakebase for AI workloads</strong>: Databricks just dropped Lakebase, a new PostgreSQL-compatible database specifically engineered for AI workloads. It aims to merge operational database capabilities with their existing data lakehouse architecture, offering a serverless OLTP solution that separates compute from storage. This is a big deal for real-time ML feature serving and AI agent memory, and it’s already GA on AWS with Azure in preview, according to Databricks. <a href="https://app.daily.dev/posts/56RHAShlB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google’s embedding model deprecation causes vendor lock-in</strong>: Google deprecated its <code>text-embedding-004</code> model, instantly bricking a production RAG-based movie recommender because different embedding models create incompatible vector spaces. This highlights a nasty vendor lock-in where your persisted embeddings are tied to a specific model’s geometry, not just an API. The takeaway: treat embedding models as infrastructure and design for change if you’re persisting embeddings at scale. <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</strong>: The Linux kernel 7.0 rc1 is out, and Linus Torvalds says the major version bump is just because he gets confused by big numbers, not a huge feature. Still, it brings cemented Rust support, faster cache clearing, non-disruptive kernel updates, and updated AMD/Intel silicon support. Plus, there are significant performance improvements for RISC-V and LoongArch, as reported by multiple sources. <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>OpenClaw gateway launches AI agent orchestration dashboard</strong>: OpenClaw Gateway’s ‘Mission Control’ is an open-source AI agent orchestration dashboard built with Next.js 14 and TypeScript, designed to create tasks, plan with AI, and manage specialized agents on a Kanban board. <a href="https://app.daily.dev/posts/l0KLAjYGc" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Uber and OpenAI reveal infrastructure traffic control platforms</strong>: Uber’s “Global Rate Limiter” and OpenAI’s “Access Engine” are new infrastructure-level traffic control platforms, with Uber’s handling 80 million requests per second using “probabilistic shedding” and OpenAI’s using a “credit waterfall” mechanism. <a href="https://app.daily.dev/posts/OVoWP2oNj" 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 local-first microVM sandbox for AI agent workloads on macOS, using Apple’s Virtualization.framework to provide secure, ephemeral, and reproducible Linux environments without Docker. <a href="https://app.daily.dev/posts/TqgsBjEoV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Timeframe e-paper dashboard project evolves with home assistant</strong>: A developer’s personal e-paper dashboard project, Timeframe, has evolved over a decade, now using a 25.3&quot; Boox Mira Pro and Home Assistant as its primary data source, with a focus on displaying actionable information. <a href="https://app.daily.dev/posts/hKJfGUQ4Z" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>VS Code extension ‘fresh file explorer’ released</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 and one-click restore for deleted files. <a href="https://app.daily.dev/posts/pQedWHBRB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Rust rwlock underperforms mutex in read-heavy workloads</strong>: A hands-on investigation found that <code>RwLock</code> performed five times worse than <code>Mutex</code> for a read-heavy LRU tensor cache in Rust due to cache-line ping-pong, 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>
<li><strong>Database design starts with consequences, not tables</strong>: Good database design starts with understanding business consequences, advocating for normalizing to 3NF, deliberate denormalization, and using constraints to embed business rules, as performance issues are often schema problems. <a href="https://app.daily.dev/posts/fI73pTtiz" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic’s claude compiler video draws criticism</strong>: Anthropic released a marketing video showing Claude 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 autonomy, not a straightforward compiler-building demo. Because, apparently, we’re all supposed to believe AI can just whip up a Linux-compiling C compiler like it’s nothing. <a href="https://app.daily.dev/posts/vGYk0DbTO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gemini 3.1 pro disappoints in coding tasks</strong>: Google’s Gemini 3.1 Pro, despite its impressive benchmark scores, apparently “underperformed” Claude Opus 4.6 in real-world coding and UI generation tasks. The article calls it “pure disappointment,” citing tool-calling failures and incomplete features. It’s almost like benchmarks don’t always translate to actual utility, who would’ve thought? <a href="https://app.daily.dev/posts/bIAFUawfa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta patents AI for deceased user simulation</strong>: Meta was granted a patent for an AI system that simulates 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 another groundbreaking feature for the metaverse. <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
