Pose Resolution Architecture

Hey everyone,

I am not a scientist but I got really intrigued by "The Thousand Brains Theory" (TBT) and some other books I came across. What started as going back and forth with Claude led to a whole bunch of experiments, a scientific approach to collaborating with Claude and, what I consider, an uncommon approach to intelligent systems with online learning capabilities.

Having a frozen llm model is one thing, but in my opinion, it will not lead to anything substantial. After all, it remains a game of playing with prompt/context/ engineering. If we want to get somewhere meaningful, we need a system that never stops learning. That's where the whole PRA idea came from and honestly, the journey link above explains it better than I can directly.

Instead of focussing on a reward-based system, PRA uses a drive and curiosity as I discovered in https://impire.io/poseres-book/part-3-the-mechanism/09-wanting-things.html.

I am not claiming anything here, just want to share what I think is interesting as many of the things sure did surprise me. All of this is done in the open and I explicitly keep a journal. The website is updated based on that journal so the "book" you see, is actually more of some sort of diary.

I can't jot down the whole journey here as it is a multi-week storyline. So take a look at the journey and the project

I hope it is useful to some of you! If not, either way is fine by me. I will just continue with this because it is fun and keeps my brain going ;)

Cheers! D.