1
Turbovec – Google's TurboQuant for vector search in Rust
内存压缩8倍还比FAISS快，免训练免调参的Rust向量索引，值得一试
Article URL: https://github.com/RyanCodrai/turbovec Comments URL: https://news.ycombinator.com/item?id=49349898 Points: 289 # Comments: 34
2
Attribute Knowledge RAG Pattern for Governed AI Agents
提出一种属性知识RAG模式，通过本地与生产环境的混合检索架构，让LLM在受控条件下选择字段，避免幻觉。
Article URL: https://superml.dev/attribute-knowledge-rag-pattern-llm-governed-attributes Comments URL: https://news.ycombinator.com/item?id=48810859 P…
3
Real-Time Hard Negative Sampling via LLM-based Clustering for Large-Scale Two-Tower Retrieval
用LLM聚类实现实时难负样本采样，直击双塔检索训练瓶颈，规模越大收益越明显。
arXiv:2607.00448v1 Announce Type: cross Abstract: The two-tower model has been widely used for large-scale recommendation systems, particularly in the…
4
Show HN: GalaxDB – an open-source AI-native database(OLTP+vector+versioning)
AI时代原生数据库来了，把向量检索、事务处理和版本管理装进一个库，一条SQL就能存文本并自动向量化
Article URL: https://github.com/zentrix-innovative-labs/galaxdb Comments URL: https://news.ycombinator.com/item?id=48708750 Points: 1 # Comments: 0
5
How does an AI agent pick from 686 skills in a second?
实测AI代理如何在一秒内从686个技能中精准选中最优解，准确率68%的秘密武器是语义路由+向量检索。
I ran an empirical test on the "skills as semantic router" pattern for Claude Code agents. I indexed 686 randomly sampled skills from a 4,556-skill co…
6
Human-Like Document AI
像人类一样理解复杂文档，无需向量数据库，精准答案可溯源验证
Article URL: https://pageindex.ai/ Comments URL: https://news.ycombinator.com/item?id=48170832 Points: 2 # Comments: 1