從競爭對手 Zilliz 的角度看 AWS 推出的 S3 Vectors

看到 Zilliz 看 AWS 新推出的 S3 Vectors:「Will Amazon S3 Vectors Kill Vector Databases—or Save Them? (via)」,算是解答了一些當初 AWS 在推出 S3 Vectors 時寫下來的問題:「關於 Embedding 搜尋問題的坑」。 看起來現有的技術是有機會搬出這樣的產品,不算魔法...

Gea-Suan Lin's BLOG
🌘 Amazon S3 Vectors 會扼殺向量資料庫,還是拯救它們?
➤ S3 Vectors:向量搜尋成本殺手,抑或生態系新夥伴?
https://zilliz.com/blog/will-amazon-s3-vectors-kill-vector-databases-or-save-them
本文探討 Amazon S3 Vectors 的出現對向量資料庫市場的影響。作者認為,S3 Vectors 雖然能大幅降低儲存成本並提供極致擴展性,尤其適合低查詢頻率且能容忍延遲的工作負載,但其效能限制和較高的查詢延遲,使其難以取代專業向量資料庫。相反地,S3 Vectors 更適合作為專業向量資料庫的補充,形成一種分層儲存的解決方案,結合兩者的優勢。
+ 這篇文章提供了非常有價值的見解。我一直擔心向量搜尋的成本問題,S3 Vectors 的出現確實讓人眼前一亮,但效能的取捨是關鍵。
+ 身為 Milvus 的使用者,我認為作者的觀點很有道理。S3 Vectors 聽起來很吸引人,但對於需要即時
#向量資料庫 #Amazon S3 #Zilliz #Milvus #AI #向量搜尋
Will Amazon S3 Vectors Kill Vector Databases—or Save Them? - Zilliz blog

AWS S3 Vectors aims for 90% cost savings for vector storage. But will it kill vectordbs like Milvus? A deep dive into costs, limits, and the future of tiered storage.

Amazon S3 #vectors are coming for your precious vector databases, like a bear in a honey warehouse 🐻🍯. But don't worry, #Zilliz is here to reassure you with a deluge of buzzwords and a pricing calculator that even your cat could use 🐱🧮. Grab a coffee, folks, this one's a real nail-biter! ☕️💤
https://zilliz.com/blog/will-amazon-s3-vectors-kill-vector-databases-or-save-them #AmazonS3 #VectorDatabases #TechBuzzwords #CloudComputing #HackerNews #ngated
Will Amazon S3 Vectors Kill Vector Databases—or Save Them? - Zilliz blog

AWS S3 Vectors aims for 90% cost savings for vector storage. But will it kill vectordbs like Milvus? A deep dive into costs, limits, and the future of tiered storage.

New Open at Intel podcast is out! I had a fantastic chat with Stephen Batifol of #Zilliz about #Milvus and the power of vector databases, as well as the MLOps community and how projects like OPEA can help developers new to #AI/ML find their way. Check it out!

Subscribe to Open at Intel in your favorite podcast app or find it at https://openatintel.podbean.com/e/understanding-milvus-the-power-of-a-vector-database/

#Podcast #NewEpisode #AI #OPEA

Understanding Milvus: The Power of a Vector Database | Open at Intel

In this episode, Steven Batifol, a Developer Advocate at Zilliz, discusses his role in fostering the MLOps community, the significance of vector databases like Milvus, and the importance of open source ecosystems. We covered the excitement of developing creative demos, the challenges facing developers in the AI space, and the rapid advancements in LLMs and AI agents. We even learn some trivia about Germany and fax machines!    00:00 Introduction 00:16 Developer Advocacy 01:02 The MLOps Community in Berlin 01:51 Joining Zilliz and Working with Milvus 04:46 Fun and Creative Demos 10:21 Challenges in the AI/ML Community 13:00 The Importance of Open Source 17:02 Upcoming Open Source Summit Presentation 20:14 Future of AI and LLMs 24:24 Conclusion   Guest: Stephen Batifol is a Developer Advocate at Zilliz. He previously worked as a Machine Learning Engineer at Wolt, where he created and worked on the ML Platform, and previously as a Data Scientist at Brevo. Stephen studied Computer Science and Artificial Intelligence. He is a founding member of the MLOps.community Berlin group, where he organizes Meetups and hackathons. He enjoys boxing and surfing.   

There are so many #vectordatabase companies now (#Chroma, #Qdrant, #Weaviate, #Pinecone, #Zilliz to name a few) that it feels like a big #AI bubble is growing. All you need for a vector database is basically an embedding model and a retrieval method based on a distance metric for the vectors like cosine similarity. In Berlin alone there are two companies - mixedbread.ai and jina.ai - who only work on embedding models. Is this an AI bubble which is going to burst or something different?
🌗 近似最近鄰居 Oh Yeah (Annoy) - Zilliz 向量資料庫部落格
➤ 近似最近鄰居 Oh Yeah (Annoy) 的工作原理和 Python 實現
https://zilliz.com/learn/approximate-nearest-neighbor-oh-yeah-ANNOY
本文介紹了近似最近鄰居 Oh Yeah (Annoy) 演算法,它使用一個樹的森林來進行最近鄰居搜尋。Annoy 是一種基於二元搜尋樹的索引結構,通過重複劃分向量空間並僅搜索部分劃分來尋找最近鄰居。本文詳細介紹了 Annoy 的工作原理並提供了 Python 實現的示例。
+ 這篇文章很清楚地解釋了 Annoy 演算法的工作原理,並提供了實際的程式碼示例,非常有幫助。
+ Annoy 看起來是一個很有用的演算法,我想嘗試在我的項目中應用它。
#近似最近鄰居 #向量資料庫 #Annoy #Zilliz
Approximate Nearest Neighbor Oh Yeah (Annoy)

Explore the efficiency and accuracy of Annoy, the game-changing approximate nearest neighbor algorithm.