<h2>Must know</h2>
<ul>
<li><strong>Databricks launches Lakebase for AI workloads</strong>: Databricks has launched Lakebase, a new PostgreSQL-compatible database specifically designed for AI workloads. It aims to bridge operational database capabilities with their data lakehouse architecture, offering a serverless OLTP solution that separates compute from storage. Lakebase supports up to 8TB per instance, includes Postgres 17 with pgvector, and integrates directly with the Databricks Data Intelligence Platform for transactional data querying without ETL. It’s currently GA on AWS, with Azure in public preview and GCP planned for later in 2026. <a href="https://app.daily.dev/posts/56RHAShlB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Google embedding model deprecation causes RAG system failure</strong>: Google deprecated its <code>text-embedding-004</code> model, rendering thousands of stored vectors unusable for a production RAG-based movie recommender. This highlights a critical vendor lock-in issue: different embedding models produce incompatible vector spaces, making migration difficult. The solution involved re-embedding the entire dataset using FastEmbed and a blue/green swap for the Qdrant collection, emphasizing the need to treat embedding models as infrastructure and design for change if 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 has been released, with Linus Torvalds humorously attributing the major version bump to his dislike of large numbers, not a significant feature milestone. This release solidifies Rust support, brings faster cache clearing, non-disruptive kernel updates, and updated AMD/Intel silicon support. It also includes performance improvements for RISC-V and LoongArch, and notably, removes a legacy 1990s IBM ThinkPad modem driver. This continues the kernel’s steady evolution, focusing on stability and hardware compatibility. <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>Kubernetes 1.35 enhances AI/ML workload support</strong>: Kubernetes v1.35, codenamed ‘Timbernetes’, introduces features like workload-aware scheduling with gang scheduling (alpha) and stable in-place Pod resource resize, strengthening its role as an operational platform for AI/ML workloads. <a href="https://app.daily.dev/posts/mAyDHBOpw" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Ubuntu 26.04 LTS to default to OpenJDK 25</strong>: Ubuntu 26.04 LTS, ‘Resolute Raccoon’, will default to OpenJDK 25, aligning LTS release cycles and bringing improvements in memory efficiency, garbage collection, and ahead-of-time profiling. <a href="https://app.daily.dev/posts/acjUVfZBa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Spring ecosystem projects hit second milestones</strong>: Second milestone releases for multiple Spring ecosystem projects, including Spring Boot 4.1.0-M2 and Spring Data 2026.0.0-M1, signal progress towards upcoming stable versions for Java enterprise applications. <a href="https://app.daily.dev/posts/4cQgIuMmg" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Developer productivity lost to non-coding tasks</strong>: A developer’s account highlights that a 10-second code fix took over 10 hours due to non-coding tasks like fighting sandboxed test environments, CUDA conflicts, and build system issues, emphasizing that LLM tools don’t help with these bottlenecks. <a href="https://app.daily.dev/posts/OBIqfWI7x" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Shuru offers local microVM sandbox for AI agents on macOS</strong>: Shuru is a new local-first microVM sandbox for macOS, using Apple’s Virtualization.framework to provide secure, ephemeral, and reproducible environments for AI agent workloads without Docker. <a href="https://app.daily.dev/posts/TqgsBjEoV" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>JDK 24 brings major security enhancements</strong>: JDK 24 introduces significant security enhancements, including post-quantum cryptography (ML-KEM, ML-DSA), core cryptography improvements, PKI updates, 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>ChatDetector uses LLMs to find RM-API misuse</strong>: ChatDetector, an LLM-powered system, detects resource management API misuse in software, identifying 115 security bugs across six popular libraries by using a ReAct-inspired framework and two-dimensional prompting. <a href="https://app.daily.dev/posts/UGo4Ea1t8" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Uber and OpenAI ‘reveal’ traffic control platforms</strong>: Uber and OpenAI have ‘revealed’ their infrastructure-level traffic control platforms, ‘Global Rate Limiter’ and ‘Access Engine’ respectively, according to InfoQ. Uber’s system handles 80 million requests per second with ‘probabilistic shedding,’ while OpenAI uses a ‘credit waterfall.’ Because, apparently, these tech giants just now decided to tell us they have systems to manage their traffic. Groundbreaking. <a href="https://app.daily.dev/posts/OVoWP2oNj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Meta patents AI to simulate deceased users</strong>: Meta was granted a patent for an AI system that simulates a deceased user’s social media activity, including posting and chatting, by training an LLM on their past data. Because nothing says ‘grief counseling’ like a chatbot pretending to be your dead relative. Ethical concerns? Mental health concerns? Nah, just another innovative way to keep users ‘engaged’ even after they’ve shuffled off this mortal coil. <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Neiry tests ‘bio-drones’ with neural interface pigeons</strong>: Neiry, a deeptech company, is testing ‘bio-drones’ – pigeons with neural implants, solar panels, and cameras – to guide them along preset routes for infrastructure inspection and search-and-rescue. They claim ‘privacy compliance’ through on-device processing. So, we’re just strapping cameras to birds and calling them ‘bio-drones’ now? And trusting a company to filter out identifiable details from pigeon-cam footage? What could possibly go wrong with brain-controlled surveillance birds? <a href="https://app.daily.dev/posts/m7v8MA0Dh" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
