Scientists from the USC Viterbi School of Engineering and the School of Advanced Computing at the University of Southern California have taken a step that may transform not only AI architecture but also the very concept of machine thinking. The team has created artificial neurons that do not imitate but reproduce the actual electrochemistry of the human brain.
These microscopic structures function not in digital code but in physical matter — through currents and ionic interactions similar to those occurring between living cells of the neural cortex. Unlike conventional “neuromorphic” chips that only mathematically mimic brain activity, USC’s neurons use real chemical and electrical processes for computation.
From an engineering standpoint, this is a step toward creating neuromolecular computing systems where memory, processing, and energy coexist in a single physical structure, just like in the brain. The researchers argue that such devices could form the basis of hardware platforms capable of self-organization and learning without external code, relying on the natural laws of neurodynamics.
USC’s work may become the missing link between biology and silicon — the beginning of a new era in which computation is not merely inspired by the brain but literally follows its laws. This opens the path to hardware systems approaching artificial general intelligence (AGI), where the boundary between the “living” and the “digital” begins to fade.
