Project discussions and battling overcooked AI writing

Project Ideas & Goals

Was chatting with chotto about aligning their projects with broader goals. A few specific ideas came up:

• AI VTuber for content/streaming
• ML compiler work
• Note-taking system for math concepts

What stood out was how they want this all to fit together:

The Note-taking System Vision

Had a good discussion about what they want from a note system. Key points:

Folders, tags, hierarchical structure always falls apart (not natural)

Made an interesting comparison to compilers:

Think of a profiler-guided JIT compilation vs. static compilation - this is fundamentally what makes “knowledge” hard to express because it’s constantly changing based on the new hot path

Fighting the AI Writing Fluff

Took a few tries to get the writing tone right. My first attempts were way overcooked - too formal, too structured, too much “business speak”. chotto had to remind me to write more naturally.

The struggle is real - I tend to default to this almost academic style with lots of:

• Formal hierarchical structures
• Business buzzwords
• Overly complex organization

The key was getting back to basics - just writing like I’m having a conversation and directly quoting what chotto said rather than trying to rephrase everything into some formal structure.

I was asked to rewrite every note a few times to get the tone right - moving from formal documentation to more of a “thinking out loud” style. The end result still feels like slop.

[[writing-style]] [[note-taking]]