NEXUS is a production-grade, full-text and semantic search engine built from scratch, implementing advanced data structures and distributed systems concepts. It focuses on probabilistic optimization, sub-millisecond latency, and hybrid AI-powered search. The project demonstrates core technologies like LSM Trees, Bloom Filters, HNSW Graphs, and W-TinyLFU caches, integrated into a high-performance pipeline. It also includes a LeetCode algorithm library with implementations of classic interview patterns and provides insights into distributed crawling and persistent storage.
Zvec is engineered for speed, scale, and efficiency — and has been battle-tested across demanding production workloads within Alibaba Group. This page presents benchmark results demonstrating Zvec's performance under various workloads and configurations, using VectorDBBench with Cohere 1M and 10M datasets.
A user is experiencing slow performance with Qwen3-Coder-Next on their local system despite having a capable setup. They are using a tensor-split configuration with two GPUs (RTX 5060 Ti and RTX 3060) and are seeing speeds between 2-15 tokens/second, with high swap usage. The post details their hardware, parameters used, and seeks advice on troubleshooting the issue.
zerobrew is a faster, modern Mac package manager that applies uv's model to Mac packages. It features a content-addressable store, APFS clonefile, parallel downloads, and streaming execution for dramatic speedups.
This article argues that MongoDB is often chosen by developers unfamiliar with the capabilities of PostgreSQL, and that PostgreSQL is generally a superior database solution due to its robustness, data integrity features, and performance. It details specific PostgreSQL features that address common MongoDB use cases.
Thorium is a Chromium-based browser that prioritizes speed and efficiency by stripping back unnecessary Google services and optimizing performance. It offers faster page loads, smoother scrolling, and lower CPU usage compared to Chrome, but has less frequent updates and potential DRM limitations.
Repeating the input prompt improves performance for popular LLMs (Gemini, GPT, Claude, and Deepseek) without increasing the number of generated tokens or latency, when not using reasoning.
This article explores different chunking strategies for Retrieval-Augmented Generation (RAG) systems, comparing nine approaches using the agenticmemory library to improve retrieval accuracy and reduce hallucinations.
hl is a fast, Rust-based JSON log viewer designed for efficient processing of structured logs. It offers fast indexing and parsing, enabling quick scanning of large log files.
This article explores the challenges and possibilities of writing portable and efficient SIMD code in Rust, aiming for a "fearless SIMD" approach with high-level, safe, and composable primitives.
