MyScale Blog
myscale.com · Company blog · on HN since 2023
MyScale, a company selling a vector database built on ClickHouse, writes about RAG, SQL plus vector search and benchmarks.
What HN saysMostly MyScale staff comment, promoting its SQL support, joint queries on structured and vector data, and its filtered-search algorithm; the comments give little outside opinion.
15HN people shared it
6linked it in comments
59stories submitted
169story points
2 years agolast shared
2023people per year2026
Most-shared posts
- 1A Comprehensive Comparison of Prompt Engineering, Finetuning and RAG3 people · 19 points · first shared 2024 · 4 comments on HN
- 2Challenges of Scaling RAG Applications3 people · 11 points · first shared 2024 · 1 comments on HN
- 3Show HN: Langchain's new member-A SQL+Vector database built on ClickHouse3 people · 7 points · first shared 2023 · 4 comments on HN
- 4When SQL Where Meets Vector Search3 people · 4 points · first shared 2023 · discussion on HN
- 5Everything You Can Do with 5M Vectors3 people · 4 points · first shared 2024 · 3 comments on HN
- 6Vector DB Benchmark (ClickHouse vs. Postgres vs. Others)3 people · 3 points · first shared 2023 · discussion on HN
- 7Hook Your Knowledge Base into GPTs2 people · 7 points · first shared 2023 · discussion on HN
- 8SQL+Vector: Empowering GenAI Applications with Relational Vector Databases2 people · 3 points · first shared 2023 · discussion on HN
- 9Understanding Vector Indexing: A Comprehensive Guide2 people · 2 points · first shared 2024 · discussion on HN
- 10A Deep Dive into SQL Vector Database2 people · 2 points · first shared 2023 · discussion on HN
- 11An In-Depth Guide to Contrastive Learning: Techniques, Models, and Applications1 people · 30 points · first shared 2024 · 4 comments on HN
- 12Why SQL for Retrieval-Augmented Generation (RAG)1 people · 16 points · first shared 2024 · discussion on HN
- 13LangChain vs. LlamaIndex1 people · 11 points · first shared 2024 · 9 comments on HN
- 14Streamlit vs. Gradio: The Ultimate Showdown for Python Dashboards1 people · 6 points · first shared 2024 · discussion on HN
- 15Don't Build Your Future on Specialized Vector Databases1 people · 6 points · first shared 2024 · discussion on HN
What HN people wrote when they linked it
lqhl · 2023-08-21
I work on MyScale (https://myscale.com), a fully-managed vector database based on ClickHouse.
Some unique features of MyScale:
1. This solution is built on ClickHouse and offers comprehensive SQL support. Our users leverage vector search for a wide range of interesting OLAP use cases.
2. We utilize a property vector search algorithm called the multi-tier tree graph (MSTG). This algorithm is significantly faster than HNSW for both vector index building and filtered vector searches.
3. We utilize NVMe SSDs for the vector index cache, which greatly reduces the cost of hosting millions of vectors.
chathero · 2023-04-14
Hey everyone,
Looking for a high-performance database that lets you use SQL to perform joint queries on both structured and unstructured data? Look no further than MyScale!
With MyScale, you can manage both structured and unstructured (vectorized) data, and perform joint queries and analytics on both types of data using SQL. Plus, MyScale's cloud-native OLAP architecture, built on top of ClickHouse, enables lightning-fast data processing even on massive datasets.
We're excited to announce that MyScale is now in beta, and we'd love to get some early feedback from the community! If you're ready to revolutionize your applications with lightning-fast results and flexible querying, check out MyScale at https://myscale.com.
coolkid0329 · 2024-03-29
Thank you for your interest and question. Simply put, you may often get unrealistic answers when trying out chatbots, i.e. the "hallucination of a large language model". The integration of RAG and MyScale effectively solves this hallucination and improves the accuracy of answers. For a concrete example, please refer to this blog:Teach your LLM to Always Answer with Facts not Fiction（https://myscale.com/blog/teach-your-llm-vector-sql/）. If you want to customize your personalized solution, please contact us（https://myscale.com/contact/） and we will give you the best price and the best quality.
Bella-Xiang · 2024-03-22
MSTG combines the best parts of both hierarchical graph and tree structures in its design.
A graph algorithm excels at initial convergence, and is usually faster at unfiltered search. However, its efficiency is severely hampered in filtered search. On the other hand, tree algorithms are slower and have lower accuracy for unfiltered search, but tree traversal is unaffected by filtered elements and retains performance for filtered search, as illustrated in the figure below. Therefore, combining these two algorithms in MSTG yields high performance and high accuracy for both cases and achieves fast index build time.
You can learn more from this blog: https://myscale.com/blog/optimizing-filtered-vector-search/
soft_camel · 2023-05-08
An alternative to Pinecone is MyScale DB. It is SQL-based vector database and has generous free tier now. https://myscale.com/
1. Manages both structured and vectorized data in a single database and can perform joint queries and analytics on both types of data.
2. Cloud-native OLAP database architecture enables operations on vectorized data to be executed with astounding speed.
3. Complete and extended SQL support for all data operations, accessible via developer tools such as Python SDK.
basil-rash · 2024-04-02
This page has a fairly helpful breakdown of the various types of indexes available, but from what I can tell it doesn’t explicitly mention the correctness trade offs made when using an index vs a pure table scan. I know in pgvector HNSW for instance is inexact. But it doesn’t really matter in a RAG context - even the embeddings themselves are stochastic.
https://myscale.com/blog/everything-about-vector-indexing/