Future Agent System
A speculative architecture for a multi-agent system that self-organises, delegates subtasks, maintains a shared memory store, and executes goals autonomously without continuous human supervision.
Self-Healing Code
A thought experiment on software that monitors its own runtime behaviour, detects logic failures, and attempts autonomous patching through constrained program synthesis.
Edge Intelligence Network
Exploring how small AI models can be distributed across edge devices, forming a mesh intelligence network that operates with no central server and no internet dependency.
Persistent Digital Memory
A design for a personal memory system that captures, indexes, and retrieves life experiences, learnings, and decisions — a second brain that runs fully locally.
Some ideas become research
When an idea is strong enough, I turn it into a proper research article with real architecture and analysis.