skanga

To be fair, these are not prompted as has been described here. These are actually "few shot" examples in the sentiment prompt. It is presented as an example under "distinguish opinion from event" but not as "do-this" evidence. Still, hard-coded positive Nvidia/NVDA examples in a generic prompt are unnecessary prompt contamination. At best this a real bias risk, but not a strong deterministic bug.

In case there is interest, I've got a fork with some custom improvements. See section "What this fork adds" in README.md

https://github.com/skanga/TradingAgents

Thanks

MDedit gives you explicit Edit / Split / Preview modes which is useful when checking what some Markdown syntax loooks like. MDedit allows exporting rendered HTML, PNG, print-to-PDF, individual tables as CSV, and diagrams as SVG. HTML can embed local assets. I did not see those in my (admittedly skimpy) review of writer.

Thanks for the feedback. Added the screenshot. Not familiar with writer-computer. Will take a look.

Thanks for being responsive. I'll try it.

Codex?

Are there any instructions on self hosting?

Fantastic Work!

You mentioned that "Everything beneath the Grace Cathedral experience is built on open source foundations." How feasible is it for a novice to build such an experience. Say I wanted one for my own house (should be MUCH less challenging)

So beautifully done, its mindblowing. Please blog about your process a bit more.

It does not work on Windows. Should be easy to add that.

Agreed. That's more interesting than the game - for some at least ...

But does it work?

What did you try? What did it do?

TDD == Prompt Engineering, for Agentic coding tasks.

I'm working on Conductor

https://github.com/skanga/Conductor

Conductor is a LLM agnostic framework for building sophisticated AI applications using a subagent architecture. It provides a robust platform for orchestrating multiple specialized AI agents to accomplish complex tasks, with features like LLM-based planning, memory persistence, and dynamic tool use.

It provides a robust and flexible platform for orchestrating multiple specialized AI agents to accomplish complex tasks. This project is inspired by the concepts outlined in "The Rise of Subagents" by Phil Schmid at https://www.philschmid.de/the-rise-of-subagents and it aims to provide a practical implementation of this powerful architectural pattern.

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.

https://github.com/skanga/JarGet is a slightly similar tool for Java

What about Tablesaw, Apache Arrow? How does this compare ...

Looks like Deepseek is having it's Tiktok moment!

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.

Thoughts?

Electricity, the printing press, etc were all privately owned.

This is probably accurate.

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.

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?

You are right. But IMO you're still missing his main point. His thesis is in that case, to rewrite ONLY THE REALLY BAD CODE.

There is not team that is 100% bad so you don't need a full rewrite of 100% of the code base.

Hi Burak, I saw cx_Oracle in the requirements.txt but the support matrix did not mention it. Does this mean Oracle is coming? Or a typo?

Found some potential bugs. Please check the github issues page.