MD-OS - Artificial Prefrontal Cortex

Can we build better AI by learning from human nature?

That is the question behind MD-OS Cortex, an experimental AI architecture I’m building in public.

In each experiment, I give Cortex a real task. It retrieves the relevant context, questions its own plan, operates software or a robot through controlled interfaces, checks what actually happened, and updates its memory only when the result is verified.

This channel documents the entire process on Linux and robotic systems: the plans, actions, mistakes, failed attempts, corrections and final evidence.

No scripted success. No hidden magic. No claims without observable results.

One new experiment every week.