I read a paper called "The Rise of Subagents" by Phil Schmid at https://www.philschmid.de/the-rise-of-subagents and thought it was an incredibly powerful architectural pattern for running AI agents with complex tasks.
So, I decided to build a practical implementation of this system with a central Orchestrator that manages a fleet of implicit or explicit Subagents. Each subagent is a specialized, isolated AI agent designed to perform a specific subtask. More details in the repo README at https://github.com/skanga/conductor
It's hard to evaluate such a tool. I scanned my OSS MCP server for databases at https://github.com/skanga/dbchat and it found 0 vulnerabilities. Now I'm wondering if my code is perfect :-) or the tool has issues!
DBChat is a powerful MCP server that lets you have natural language conversations with your database from clients like Claude Desktop. Ask it to do complex analysis, generate beautiful visualizations, or build custom interactive dashboards based your data. Works with any JDBC-compatible database with support for most SQL DBs like PostgreSQL, MySQL, Oracle, SQL Server, SQLite, MongoDB, etc.
Good question. I’ll publish benchmarks soon, but the core difference is that Fahmatrix is fully Java, no JNI, and minimalistic — ideal for small projects or environments like Android. Tablesaw and Arrow are more powerful, but heavier. Fahmatrix aims to be the “just enough” middle ground.
Hi Maxwell, I did work on some similar stuff and have some thoughts, suggestions & connections that could help. Feel free to reach out to my yc username at googles email service.
Sidebar: interesting how you typed out the email address (I’ve also seen variations like me at gmail dot com). This worked for spam and phishing crawlers that just did a brute force search across texts. But now that we have LLMs, I’m wondering if this will be effective anymore, since it gets the context as well and is not just a “dumb” regexp search?
But GenAI is a major upheaval. In most those, a massive amount of initial capital is invested, but the payoff happens slowly - only over decades. But the payoffs are huge.
Think about electricity. Building out the grid was HUGELY expensive. But then benefits are derived for decades. Same for the highway system, railway system, etc.
Not sure public infrastructure is a good comparison, because that kind of stuff benefits the public.
GenAI is definitely a major upheaval, and the payoffs will definitely be huge, but those payoffs will go to those who fund it at the detriment of the working class who haven’t been sharing in the productivity gains made by capital.
I just don’t see the public benefit that we see from public infrastructure. If anything GenAI will exacerbate and accelerate all the structural issues we have in our economy and society.
You mentioned that spRAG uses OpenAI for embeddings, Claude 3 Haiku for AutoContext, and Cohere for reranking. Can you explain why & how did you make those choices?
Those are just the defaults, and spRAG is designed to be flexible in terms of the models you can use with it. For AutoContext (which is just a summarization task) Haiku offers a great balance of price and performance. Llama 3-8B would also be a great choice there, especially if you want something you can run locally. For reranking, the Cohere v3 reranker is by far the best performer on the market right now. And for embeddings, it's really a toss-up between OpenAI, Cohere, and Voyage.
I bet you'll get a lot more adoption if you put info about using it with local self-hosted LLMs there. I'll never trust a cloud service with the documents I want to RAG.
So, I decided to build a practical implementation of this system with a central Orchestrator that manages a fleet of implicit or explicit Subagents. Each subagent is a specialized, isolated AI agent designed to perform a specific subtask. More details in the repo README at https://github.com/skanga/conductor
reply