Tag
Meta 分享其“薄编排器”架构实践：用一个 1300 行、零领域逻辑的路由器把 20+ 工作流分发给 14 个专业化子代理，显著提升了数据管线代理在生产环境中的准确性与效率。
A developer-built open-source Rust CLI/TUI called repOx that packs codebases into LLM prompts in under 15ms, stripping lockfiles, binaries, and artifacts while supporting token budgets for major model contexts.
RepoLens is a zero-dependency Go tool that packs local codebases into a single file for LLM context, featuring an interactive TUI file explorer, live token counting, comment stripping, and secret scanning.
EverMind team under Shanda Group founder Chen Tianqiao has open-sourced EverMemOS (EverOS), a brain-inspired intelligent memory operating system designed to solve the forgetting and context-limited problems of large language models by simulating human memory mechanisms, providing long-term memory infrastructure for AI Agents.
Aravind Jayendran argues that while longer context windows improve LLM performance, they cannot fully replace weight updates, framing context as transient software and weights as hardware that fundamentally alters model capabilities.