We need transformer based architecture for it's attention, to interpret continuous data.

The state is the persona of the AGI and reinforcement learning is for the neuroplasticity, to learn, unlearn and evolve