Stories by marcelmarais

An MCP Server for Obsidian

Vote for the Best LLM Framework

Agents don't like your modular code

A simple engine for converting any file to Markdown

The Modern Document Processing Stack

Slightly Better LinkedIn Analytics

A verbose explanation of LLM decoding parameters

LLM Framework Leaderboard

The Straddled Matrix: improving the speed of convergence in Autoencoders

Structured Generation Shoot-Out

Q*: Improving Multi-Step Reasoning for LLMs with Deliberative Planning

Show HN: I created a free minimalist tool that sends articles to your Kindle

NumPy 2.0 Is Released

Everyone seems to have forgotten about Gemma