<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. It aims to merge operational database capabilities with their existing data lakehouse architecture, offering a serverless OLTP solution that separates compute from storage. Lakebase, which integrates technology from Neon and Mooncake Labs, supports up to 8TB per instance and includes Postgres 17 with pgvector support, making it a significant play for real-time ML feature serving and AI agent memory. <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 recently deprecated its <code>text-embedding-004</code> model, rendering existing vectors unusable for a production RAG-based movie recommender. This highlights a critical issue: different embedding models create incompatible vector spaces, leading to a hidden form of vendor lock-in where stored data is tied to a specific model’s geometry. The incident forced a full re-embedding of the dataset and a blue/green deployment, underscoring the need to treat embedding models as infrastructure and design for change. <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 and performance boosts</strong>: Linus Torvalds has released Linux kernel 7.0 rc1, a version bump he attributes to simply disliking large numbers, not a major feature milestone. Despite the humble naming, this release cements Rust support, brings faster cache clearing, non-disruptive kernel updates, and updated AMD/Intel silicon support. It also includes significant IO_uring latency fixes and memory reclamation improvements, alongside RISC-V and LoongArch performance enhancements. <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>Rust RwLock can be slower than Mutex for read-heavy workloads</strong>: A recent investigation found that <code>RwLock</code> can perform significantly worse than <code>Mutex</code> for read-heavy LRU tensor caches 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>
<li><strong>Go’s FFI now faster than GDScript and Rust</strong>: Go bindings for Godot, <code>graphics.gd</code>, have achieved FFI performance surpassing GDScript and even ‘safe’ Rust in benchmarks, thanks to call batching and Go 1.26’s cgo overhead reduction. <a href="https://app.daily.dev/posts/j8AoMYJTv" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>JDK 24 brings post-quantum cryptography and security manager removal</strong>: JDK 24 introduces significant security enhancements, including post-quantum cryptography support (ML-KEM, ML-DSA), SHA-3 performance gains, and the permanent removal of the Security Manager. <a href="https://app.daily.dev/posts/TLN02dr7b" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Shuru offers local-first microVM sandbox for AI agents on macOS</strong>: Shuru is a new macOS tool providing secure, ephemeral microVM sandboxes for AI agent workloads, using Apple’s Virtualization.framework without Docker, and featuring checkpoint/restore functionality. <a href="https://app.daily.dev/posts/TqgsBjEoV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New VS Code extension ‘Fresh File Explorer’ focuses on recent files</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>PHP implicit type casting can be risky</strong>: Relying on PHP’s implicit type casting with arithmetic tricks like <code>+ 0</code> for string-to-integer or <code>. ''</code> for integer-to-string conversions can lead to warnings and is less safe than explicit casting. <a href="https://app.daily.dev/posts/6Owe3BNN0" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Symfony weekly update includes memory and transaction fixes</strong>: The latest Symfony weekly update highlights 74 merged pull requests, including an LRU cache for <code>CssSelectorConverter</code> to prevent memory exhaustion and a fix for transaction poisoning in <code>DoctrineDbalAdapter</code>. <a href="https://app.daily.dev/posts/1bJAzjFxX" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic’s Claude compiler video is ‘dishonest framing’</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 extended autonomous operation, not a straightforward demonstration of compiler-building prowess. So, another day, another AI company trying to make their chatbot sound like a wizard, when it’s probably just a very persistent intern. <a href="https://app.daily.dev/posts/vGYk0DbTO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>AI agents are the new ‘productivity’ bottleneck</strong>: OpenClaw, an autonomous AI agent tool, is being hyped for automating everything from React dependency maintenance to compiling tech research briefings. But here’s the kicker: it can’t autonomously respond to Claude Code permission prompts, meaning you either pre-configure everything or use a <code>dangerously-skip-permissions</code> flag. So, we’ve built AI agents to automate our jobs, only to find out we still need to babysit them or just turn off the safety features. Progress! <a href="https://app.daily.dev/posts/WCj2dFu7Q" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta patents AI to simulate dead users</strong>: Meta was granted a patent for an AI system that simulates a user’s social media activity, including posting and chatting, using an LLM trained on their past data, specifically for when a user is ‘absent’ from the platform. Because nothing says ‘we care about your mental well-being’ like a digital zombie version of your loved ones posting on Facebook. Ethical concerns? Mental health concerns? Nah, just another patent for ‘engagement.’ <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
