Building Toward Mind - An Infrastructure Engineer's Exploration of AI Cognition
What happens when you take the question “could LLM inference be a cognitive primitive?” seriously enough to build something? The observation that started this I’ve always been curious about how systems work - including my own. Inspecting my systems, I noticed something about my own thinking: I don’t choose which thoughts appear. Thoughts arise (from context, memory, association, whatever the underlying machinery is) and then I align with them, steer them, or let them pass. The “self” that feels like it’s doing the thinking is really more akin to an observer or editor of a process that’s largely opaque to me. ...