AI LABS

GOTO Conferences

Jendrik Illner

Kotlin Weekly

The Linux Cast

Lobsters

<h2>Must know</h2>
<ul>
<li><strong>Google’s embedding model deprecation causes vendor lock-in</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 author, who had to re-embed their entire dataset, advises treating embedding models as infrastructure and designing 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>Databricks launches lakebase for AI workloads</strong>: Databricks just launched Lakebase, a new PostgreSQL-compatible database specifically designed for AI workloads. It aims to merge operational database capabilities with their existing data lakehouse architecture, offering a serverless OLTP database with features like instant data branching and unified access controls. This move by Databricks, leveraging tech from Neon and Mooncake Labs, signals a significant push to integrate transactional data directly into AI platforms without the usual ETL headaches. <a href="https://app.daily.dev/posts/56RHAShlB" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Linux kernel 7.0 rc1 released, torvalds muses on succession</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 overhaul. Still, it brings cemented Rust support, faster cache clearing, non-disruptive kernel updates, and updated AMD/Intel silicon support. Torvalds also mused about a future maintainer, hinting at the community’s recently developed succession plan, 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>Non-coding tasks kill developer productivity</strong>: A developer’s 10-second code fix took 10 hours due to non-coding tasks like fighting sandboxed environments, CUDA conflicts, and build failures, highlighting that LLM tools don’t help with these systemic issues. <a href="https://app.daily.dev/posts/OBIqfWI7x" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Uber and OpenAI unveil traffic control platforms</strong>: Uber and OpenAI revealed their infrastructure-level traffic control platforms, “Global Rate Limiter” and “Access Engine” respectively, designed to manage massive request volumes and ensure fair access. <a href="https://app.daily.dev/posts/OVoWP2oNj" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Fresh file explorer extension for VS Code</strong>: A new VS Code extension, ‘Fresh File Explorer,’ replaces the standard file explorer to show 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>RwLock can be slower than mutex for read-heavy workloads</strong>: Rust’s <code>RwLock</code> performed 5x worse than <code>Mutex</code> in a read-heavy LRU cache 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>Microsoft project silica advances glass storage</strong>: Microsoft’s Project Silica achieved a breakthrough in glass-based data storage, extending it to common borosilicate glass and improving writing/reading speeds, with research now complete and findings public. <a href="https://app.daily.dev/posts/KqLBpGspa" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Quantum error correction breakthrough with lattice surgery</strong>: ETH Zurich researchers made a breakthrough in quantum error correction using ‘lattice surgery,’ which measures relationships between quantum data blocks to detect errors without collapsing wave functions. <a href="https://app.daily.dev/posts/t1NKbgnH0" 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 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>
</ul>
<h2>Bullshit police</h2>
<ul>
<li><strong>Anthropic’s claude compiler video draws criticism</strong>: Anthropic released a marketing video showing Claude autonomously building a C compiler capable of compiling Linux and Doom. Critics, however, are calling out the “dishonest framing,” suggesting the real goal was merely a stress test of Claude’s ability to run for an extended period, not a straightforward demonstration of compiler-building prowess. Apparently, even AI marketing needs a reality check. <a href="https://app.daily.dev/posts/vGYk0DbTO" target="_blank" rel="noopener nofollow">Read more</a></li>
<li><strong>Gemini 3.1 pro underperforms in real-world coding</strong>: Google’s Gemini 3.1 Pro, despite its impressive benchmark scores, reportedly underperformed Claude Opus 4.6 in real-world coding and UI generation tasks, with issues like tool-calling failures and incomplete feature implementation. It seems benchmarks are great for bragging rights, but when it comes to actually getting things done, Gemini 3.1 Pro was a “pure disappointment.” Who knew real-world performance mattered more than a number? <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, by training an LLM on their past data. Because nothing says ‘grief counseling’ like a chatbot pretending to be your dead aunt. The ethical and mental health implications are, naturally, “significant.” What could possibly go wrong? <a href="https://app.daily.dev/posts/AeW0pHCcd" target="_blank" rel="noopener nofollow">Read more</a></li>
</ul>
