In order to learn and experiment with Agentic AI, I build a platform called "iTeri" which runs my "ThoughtBeat" agentic AI architecture.
During regular intervals, an LLM is asked to "Perceive", "Think" and "Act" upon its world.
This approach has proven exceedingly effective in expanding LLM autonomy and is now being used within a number of my client projects.
After having built my "ThoughtBeat" agentic AI architecture, I began to experiment with running multiple (smaller) LLMs in parallel across across multiple functional layers.
Rather than considering them as a "swarm of agents" I wanted to each LLM to act more like a "cortex" within a single "hive mind" with a shared global workspace and common set of goals.
The approach has been extremely interesting and very effective. Research is still ongoing.
I've been a big fan of "operator-style" AI projects for a while, and decided to build one myself as a Chrome Extension.
By feeding an agentic AI with a cut-down and annotated version of the browser's DOM, it was immediately (with no training) able to perform a wide range of tasks from clicking links, to filling out forms.
The potential for "client-side" agentic AI is immense and my experimentation within this space is still ongoing.
Virtual reality has always been a passion of mine, and I have built numerous VR worlds using Unity and WebXR.
My ThoughtBeat architecture is a perfect fit for autonomous agents working within a changing environment; and so I decided to tie my two passions together and built AI + VR.
Though it ints early stages, I am already having great success introducing autonomous agents to VR worlds.
During my time as a student, I wrote a number of papers and essays primarily focused on Complexity science and AI.
My particular passion was (and still is) the subject of emergence, and the study of how complex behaviours can arise from simple rules.
