<h2>Must know</h2>
<ul>
<li><strong>Databricks launches new database for AI workloads</strong>: Databricks just dropped Lakebase, a PostgreSQL-compatible database built specifically for AI workloads. It’s designed to merge operational database capabilities with their existing data lakehouse architecture, offering serverless OLTP with compute-storage separation. Currently available on AWS, with Azure in public preview and GCP coming later this year, as reported by 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, effectively bricking thousands of stored vectors for a production RAG system. The core issue is that different embedding models create incompatible vector spaces, making migration a nightmare and highlighting a hidden form of vendor lock-in. This means if you’re persisting embeddings at scale, you need to treat the model as infrastructure and plan for inevitable changes. <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 succession commentary</strong>: Linus Torvalds released Linux kernel 7.0 rc1, noting the version bump is purely cosmetic due to his aversion to large numbers, not a major feature milestone. The release includes cemented Rust support, faster cache clearing, non-disruptive kernel updates, and updated AMD/Intel silicon support. Torvalds also mused about a future maintainer, referencing the community’s recently developed succession plan. <a href="https://app.daily.dev/posts/cYP8biu57" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Good to know</h2>
<ul>
<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>Rust’s RwLock can be slower than Mutex for read-heavy workloads</strong>: A hands-on investigation found <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>
<li><strong>New VS Code extension for file navigation</strong>: Fresh File Explorer is a new VS Code extension that replaces the standard file explorer, focusing on recently modified files with Git history integration, color-coded recency heatmaps, and a ‘Resurrect’ feature for deleted files. <a href="https://app.daily.dev/posts/pQedWHBRB" 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 using Apple’s Virtualization.framework to create secure, ephemeral Linux microVMs for AI agent workloads, offering features like checkpoint/restore and opt-in networking without Docker. <a href="https://app.daily.dev/posts/TqgsBjEoV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>ChatDetector uses LLMs to find RM-API misuse</strong>: ChatDetector is an LLM-powered system designed to detect resource management API misuse in software, using a ReAct-inspired framework and two-dimensional prompting to identify and ethically report security bugs. <a href="https://app.daily.dev/posts/UGo4Ea1t8" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>New real-time glint particle rendering technique</strong>: Adobe Research, NVIDIA, and Aalto University developed a new technique for real-time rendering of millions of microscopic glint particles at over 280 FPS on consumer GPUs, generating particle positions on-the-fly to reduce memory requirements. <a href="https://app.daily.dev/posts/rT496vKCA" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Gemini 3.1 Pro: pure disappointment in coding tasks</strong>: Apparently, Google’s Gemini 3.1 Pro, despite its ‘strong benchmark scores,’ consistently underperformed Claude Opus 4.6 in real-world coding and UI generation tasks. Who would’ve thought benchmarks don’t always translate to actual utility? The unnamed ‘hands-on comparison’ even called it ‘pure disappointment.’ <a href="https://app.daily.dev/posts/bIAFUawfa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Anthropic’s compiler video: more marketing than reality</strong>: Anthropic released a marketing video showing Claude ‘autonomously’ building a C compiler capable of compiling Linux and Doom. Critics are calling it ‘dishonest framing,’ suggesting the real goal was just a stress test of Claude’s autonomy, not a straightforward compiler-building demo. Shocking, a marketing video might be overstating things. <a href="https://app.daily.dev/posts/vGYk0DbTO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta patents AI to post after you’re dead</strong>: Meta was granted a patent for an AI system that simulates deceased users’ social media activity, posting and chatting based on past data. Because nothing says ‘grief counseling’ like your dead grandma’s AI posting about her favorite cat memes. Ethical and mental health concerns? Nah, just more ‘engagement.’ <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
