shot myself in the foot by having filter_html: true in my Redcarpet setup. It was a very easy fix. #llapp

fixed a huge error #llapp

All the prompts #llapp

In the final stretch of #llapp

Early morning coding ftw #llapp

Today, we deploy. #llapp

Figured out that the LLM doesn't have enough context of the situation to do what I'm asking it to do. Now I have to figure out how to solve the problem anyway. #llapp

Cranking that OpenAI bill. $0.08!! #llapp

Working on LLM integration and file manipulation #llapp

Status report for #llapp

docx conversion day 4. #llapp

Converting docx to markdown #llapp

Fixing chunk size and retrieval. #llapp

Making a task to convert docx to md. #llapp

Product owner just changed all of the requirements. Yay! #llapp

Tweaking the RAG for better output. #llapp

Chat bot ftw #llapp I guess I'm a little late building a chat bot, but I have a really good set of unique data to use, which makes it interesting at least.

Product owner: that looks like generic content from an LLM. Me: I see the app gathering the data from the RAG... It should be working Narrator: it was not working. He had never implemented adding the RAG data to the context #llapp

RPI - Research Plan Implement. #llapp

I have a list of bugs this long! #llapp