A hybrid system pairs classical AI for routine tasks with a quantum-based conscious module for novel situations. When robots hit complex problems, the conscious module makes decisions and generates training data that updates the classical models in real time, creating a feedback loop that improves performance without waiting for external datasets.
I don't think we're going to innovate in any classical neural nets anytime soon, but there might be innovations in how to connect classical AI into this quantum conscious decision-making system. So, one of the things I'm really interested in is how do you get the best of both worlds? So, we trained all these classical AI models to do unconscious things really well from training data. But what if the consciousness module could actually generate that training data in the first place? When the robot suddenly finds itself in a completely novel situation, very complex, very dynamic, it switches into conscious decision-making mode and then switches over to the cloud platform I'm offering. It starts like making decisions tentatively, finding its way around, and it's throwing off training data at the same time. That can now be improving or training the classical AI model. So then the next time it's in that environment, it can just use the unconscious module.
