Dom Couldwell

How can you use Gen AI to put an old head on young shoulders? TL;DR it's about your AI Stack, Culture and Trust More thoughts from me here - innovationnewsnetwork.com/generative-ai-… via @InnoNewsNetwork

@heyruchir Astra from DataStax is very complimentary to the GCP stack and provides a serverless PAYG vector database solution. You can find more on the partnership here - datastax.com/blog/datastax-…

Does Google Cloud have Vector search / vector database capabilities? Some of the tasks I'm trying to achieve require creating embeddings of millions of tokens, which is too costly with OpenAI. #GoogleCloud #OpenAI #LLM

New customer story ➡️ Learn how LGBTQ+ social network and dating app Hornet is improving user experiences with AI and vector search. Learn more on the DataStax blog: ow.ly/Pm8950Q5LCS #VectorSearch #AIApplications

Check us out at GitHub Universe… @DataStax is the first vector database provider to integrate directly into the GitHub Copilot experience. Big thanks to @ShuyinZhao @rakyll for making this possible! #GitHubUniverse #GitHubCopilot #DataStax #AstraDB

@raoufdevrel I think you should! If your vector store is 'just' a vector store you're not setting yourself up to scale both in terms of performance and also use cases e.g. hybrid search. More info here on how we did this for datastax.com/products/datas…, dzone.com/articles/5-har…

Question for AI Engineers. Do you care which vector store to use in your RAG apps?

@louis030195 If you're looking for a hosted NoSQL + Vector solution check out datastax.com/products/datas…. Serverless so you only pay for what you use, generous free tier and scales to whatever you need, more here - dzone.com/articles/5-har…

Vector DB, NoSQL, or SQL? Curious if you use one only, two together, or all three? My experience - SQL DBs that do vector too are not good enough - vector DBs that let you store normal data are not good enough - never tried NoSQL + vector yet Even at small scale 10-20k users

@aiexplorations Have you tried out datastax.com/products/datas…? Generous free tier, based by Cassandra so scales to whatever you need but is also serverless so only pay for what you use. More background here dzone.com/articles/5-har…

Is Qdrant the best vector database out there? Genuine question for those who have used more than one. I for one have found Elastic really nice because it comes in cloud-based PaaS flavours, but then it isn't actually a dedicated vector DB.

While SF politician @AaronPeskin keeps trying to ban autonomous vehicles, Swiss Re—one of the world’s largest reinsurers—benchmarked Waymo’s data against 125B miles of human driving. tldr, AVs *much* safer and easier to insure than humans: > at least 300% safer than human drivers for bodily damage > at least 400% safer than humans for property damage swissre.com/reinsurance/pr…

@Higgs_X_Boson What kind of level of guidance are you looking for? Here's a CTO level guide thestack.technology/generative-ai-… and here's a customer pov datastax.com/enterprise-suc…

Anyone has good resources with working with Vector DB’s especially in production!

@ItsKrisKriz Have you tried using Astra? Here's a couple of starter guides that might be useful docs.datastax.com/en/astra-serve… docs.datastax.com/en/astra-serve…

Ok never mind the previous post... I've spent all morning trying to make a simple vector store for text embeddings work and still haven't got it. Defaulting back to Python for this one 😅

@mario_meissner try datastax.com for more background

@domlovesai No landing page?

What database should you be choosing for your new LLM project? Consider whether you really need a vector store, and also keep other factors in mind such as full-text-search. I write all about it in my new blog post, check it out here! mariomeissner.dev/posts/picking-…

What is #LangChain and what is CassIO? Discover how connecting the two can help developers tap into the power of Cassandra for LLM applications. via DataStax's Charna Parkey Watch more ➡️ ow.ly/EUf7104M5Pi

@aMichaelLyons astra.datastax.com! Generous free tier, kick ass integrations and limitless scale

What's your favorite vector DB?

Some thoughts on fine-tuning versus RAG (tl;dr personalized AI doesn't get solved by fine-tuning) thenewstack.io/fine-tuning-is… via @thenewstack

LangChain ➕ Astra DB = a rock-solid foundation for building generative AI applications. Check out what developers are building with LangChain's GenAI framework and DataStax's vector database. #VectorDB #LangChain #LLMs dtsx.io/46wSgeX

@this_is_tckb @nileshtrivedi @cvkmohan @elixirlang @elixirphoenix @josevalim @qdrant_engine @weaviate_io @langchain The developer is and always has been king. Key to their productivity is ease of use and meeting them where they are. That's why we invested in the JSON API for astra.datstax.com to make developers lives easier - datastax.com/blog/introduci…

AI/LLM ecosystem is quite dynamic and is evolving every day - python is an easy barrier and conventionally has been first choice for ML. LC has a pretty bad support in Java too. @elixirlang has a pretty strong community For eg: Pinecone Vdb has no official library but here is the community lib :hex.pm/packages/pinec… QDrant: hex.pm/packages/qdrant Weaviate: hex.pm/packages/noizu… ChromaDB: hex.pm/packages/chroma Faiss: hex.pm/packages/hnswl… and more..

I think we need a new label: Deepstack engineers. These are people who have all 3 skills: frontend/UI, backend skills (APIs/databases/systems), and Machine Learning skills. "Fullstack" has been claimed by people with only the first two skills. And when the whole construct of "apps" becomes redundant (for example: AI generates UI on the fly), the first and last may coalesce into a single role. It will be interesting to see which programming language becomes the dominant choice for this new role. Currently, Javascript owns UI, Python owns ML.

OK, I know the title is a little cheesy but what do you think? techinformed.com/rag-to-riches-…

.@DataStax 🤝🏽 @langchain some helpful resources 👇🏽 how to integrate Astra DB with LangChain for LLM workflows: youtube.com/watch?v=VzB4uq… how to build a powerful PDF Document-based Q&A System using RAG: youtube.com/watch?v=wOAV3H…

💎 Cassandra Vector Store 💎 Cassandra DB is a battle-tested, highly-available database designed to handle massive amounts of data. You can now use it as a vectorstore in @langchain 🦜🔗 JS/TS 0.0.167! Thank you alexleventer and the @DataStax team! js.langchain.com/docs/modules/d…