Milvus Blog

Bit of self-promotion, but Milvus (https://milvus.io) is another open-source vector database option as well (I can guess as to why it isn't listed in a16z's blog post). We also have milvus-lite, a pip installable package that uses the same API, for folks who don't want to stand up a local service. pip install milvus Other than that, it's great to see the shout-out for Vespa.

As mentioned in the article, I recommend Milvus (https://milvus.io) - it's open source and cloud native with standalone versions available. Alternatively, if you're looking for an open-source solution for generating embeddings, I recommend (https://github.com/towhee-io/towhee).

Congratulations. What do you think you milvus? https://milvus.io/. The difference seems significant from the architecture perspective.

Great conceptual explanation of the different primitives in LangChain. Some possible inspiration for your upcoming piece involving vector stores with LangChain - https://milvus.io/blog/conversational-memory-in-langchain.md

anybody have thoughts on this: https://engineering.fb.com/2017/03/29/data-infrastructure/faiss-a-libr… for doing the similarity search? I've also be thinking about looking into this: https://milvus.io/

https://vector.dev/ Edit: i think i mean https://milvus.io/ https://milvus.io/docs/image_similarity_search.md

Suggest you an open source project that can use the picture of the wood to search for similar wood. https://milvus.io/

Milvus[0] would be another interesting comparison. [0] https://milvus.io/

There was a similar situation in a project I worked on, at the time FAISS/Annoy/SPTAG were considered but we were really looking for something more managed, think ElasticSearch rather than Lucene. As luck would have it we stumbled upon https://www.milvus.io/ and it certainly checks a lot of boxes. But I'm rather excited to see some competition in this space, it's a broadly applicable trick so I'll definitely take a look at Vald next time it comes up as well!

The linked article about getting LLMs to critique each others' code review[1], the magpie tool[2], and also this recent article from Cloudflare about their code review stack[3] are all quite compelling. I'm fairly AI-skeptical not on grounds of "do they work" but "are they good for the world". I feel that getting AIs to do this kind of review work is a rare case that doesn't outsource thinking and deskill workers. It doesn't trigger the same alarm bells as having the AI write the code (including having the AI fix the issues it discovers). That's setting aside environmental and other ethical concerns, which are still significant to me. I have been impressed by the recent quality of AI code reviews*, but the experience of interacting with 3 separate AI reviewers via GitHub PRs is pretty terrible. Having more local-oriented and jj/rebase-aware review rounds would be great. *context: fairly large PHP/Laravel backend and Vue frontend [1]: https://milvus.io/blog/ai-code-review-gets-better-when-models-debate-c… [2]: https://github.com/liliu-z/magpie [3]: https://blog.cloudflare.com/ai-code-review/

Is it different/better than https://milvus.io ?

See https://milvus.io/

I remember thinking the same thing and this articles goes over most of the arguments here - https://milvus.io/blog/why-im-against-claude-codes-grep-only-retrieval… When it came out and I think it was Boris from Anthropic that said they experimented a lot with Vector Search and grep just worked better. You can try it out using the Claude-Context MCP - https://github.com/zilliztech/claude-context

There's two blog posts that go with this, check it out: https://milvus.io/blog/i-built-a-deep-research-with-open-source-so-can… https://milvus.io/blog/introduce-deepsearcher-a-local-open-source-deep…

Inkeep has been very solid for us and running in our Discord. https://milvus.io/community Robert spoke at our meetup and is awesome. https://www.youtube.com/watch?v=35JdjmiDvWI

I'm very sure milvus supports it. Milvus calls it hybrid search though: https://milvus.io/docs/hybridsearch.md#Conduct-a-hybrid-vector-search

Learn faiss first. Mullvus is very complicated in comparison and faiss will work in many cases. I mean look at their system diagram https://milvus.io/static/0bc2e74d0a1b20bbfb91bdbd03f77e5e/bbbf7/archit… Not fun...

Try this one maybe? https://milvus.io/docs/v2.1.x/overview.md

Some more Open Source vector search engines for ANN: https://milvus.io/ https://vespa.ai/ https://vald.vdaas.org/ https://github.com/semi-technologies/weaviate https://opendistro.github.io/for-elasticsearch/features/knn.html

Yes on online learning, as I gather from the comparison at https://milvus.io/docs/v0.6.0/about_milvus/vector_db.md

First off, I don't think there is anything wrong with MinIO closing down its open source. There are simply too many people globally who use open source without being willing to pay for it. I started testing various alternatives a few months ago, and I still believe RustFS will emerge as the winner after MinIO's exit. I evaluated Garage, SeaweedFS, Ceph, and RustFS. Here are my conclusions: 1. RustFS and SeaweedFS are the fastest in the object storage field. 2. The installation for Garage and SeaweedFS is more complex compared to RustFS. 3. The RustFS console is the most convenient and user-friendly. 4. Ceph is too difficult to use; I wouldn't dare deploy it without a deep understanding of the source code. Although many people criticize RustFS, suggesting its CLA might be "bait," I don't think such a requirement is excessive for open source software, as it helps mitigate their own legal risks. Furthermore, Milvus gave RustFS a very high official evaluation. Based on technical benchmarks and other aspects, I believe RustFS will ultimately win. https://milvus.io/blog/evaluating-rustfs-as-a-viable-s3-compatible-obj…

What about DiskANN? https://milvus.io/docs/disk_index.md

This article depicts a perfect world and links it to a solution which is fairly distant from that. I understand the wishful thinking of having a "magic box" for search infrastructure but as someone worked on web-scale search at Google for years I'd say the reality isn't that simple. 1. The real problem in embedding data lifecycle management is changing the embedding mode, which involves a migration process. You can't really solve that by simply streamline the vectorization and suddenly use a new model for new data ingested. You need the non-fancy migration process: create a new collection, batch generate new vectors with the new model, port all of them there, meanwhile doing dual write for all newly ingested documents, and switch search traffic to the new collection once batch ingestion is done. Streamlining vectorization as part of the ingestion call doesn't solve that. Though it is an interesting feature to lower mental complexity, that's why at Zilliz (a vector db startup) our product https://zilliz.com/zilliz-cloud-pipelines supports that and our open-source Milvus plans supporting streamlining API call to embedding service in 3.0 version: https://milvus.io/docs/roadmap.md. That said I must state that changing the embedding model is more difficult than what the article makes it feels like. We provide tools like bulk import to batch port a whole dataset of vector embeddings with other metadata like original text or image urls. But solving the problem with one "magic box"…

I'd be curious how they implement updating. AFAICT this is the thorniest part of working with existing open source solutions. When working with ANNOY in the past I've had data small enough to be able to recompute the full index in the background every few seconds in a background process and then swap in the built index to the process serving similarity queries. (you can see the VERY "research quality" code on Github, here's a decent starting place https://github.com/hyperstudio/spectacles/blob/master/spectacles/nndb/… EDIT: from the insertion docs https://milvus.io/docs/guides/milvus_operation.md#Insert-vectors-into-… it seems that they still ask you to re-build your indices after you insert vectors, although in some cases they can tell that they need to re-build the indices for you. Looks like the major value adds here are potentially shifting computation to the GPU and building multiple indices. I'll certainly evaluate this next time I'm building a project around vector search.