I had the same “wow” thought when found out about Graph Databases in the context of social networks.

Neo4j has vector similarity search on nodes, if it is powered to a AI RAG system we can generate nice information, content, recommendations (but not overuse it like instagram or tiktok do, since nostr doesn’t want to be like them).

Also the relationship “feature” between tables is upgraded on whole new level in Graph Databases compared with traditional sql databases like Oracle where you have to do joins on a middle table to achieve same thing.

Reply to this note

Please Login to reply.

Discussion

I created an MCP server for AI agents to create a connected knowledge graph. Neo4J is pretty cool. https://github.com/knowall-ai/mcp-neo4j-agent-memory

Wanna keep working on it, we have a hackathon coming up with large prizes for stuff like this.

So much could be done to improve it such as a better semantic search (using AI sentence transformers) or considering how agents can use that to re-organise and strength memories (akin to what we think happens we we are sleeping).

What are the details of the hackathon? In person thing?

No, online. Starts on Nov, ends in may. Similar to the one we did about heath care earlier this year.

Sounds cool. How do I find out more?

Coming soon. We are still finishing the website

Keep us posted :-)