KAG aims to make full use of the advantages of Knowledge Graph and vector retrieval, and bi-directionally enhance large language models and knowledge graphs through four aspects to solve RAG challenges: (1) LLM-friendly knowledge representation, (2) Knowledge Graph and original text The mutual index between fragments, (3) a hybrid reasoning engine guided by logical forms, (4) knowledge alignment with semantic reasoning.
The author believes that SQL is the core competency of Data Engineering. Now that ChatGPT can translate natural language into good SQLs. So Data Engineering is going to be retired?
I think it will change or alleviate the tedium of the boring bits of exploring a dataset, figuring out if the dataset can join to another dataset in the way you thought of first, and writing a few of the dashboards.
Hopefully it means I can dig into the real problem (correctness, idempotency, and performance) rather than the tedious implementation.
Maybe we should call it Quasi-real time. We get the data from the GitHub API[1], and GitHub delays the public events feed by five minutes. So the PR you see on the page actually occurred at least five minutes ago.
SD is so popular these days. It has been on the top of GitHub trending for a long time, last hour, last day, and even last week. https://ossinsight.io/#trending-repos
Is it because it represents that AI can do things that humans thought would never be replaced by AI before?
In fact, when I first opened HN, I couldn't read the comments very well. Because the font is too small, the layout is dense, and there are no pictures. But when I started reading the comments carefully, I was really surprised by so many people being active on it and so insightful most of the comments are.
Now I'm actually getting used to reading these text-only pages.
