Researchers at the Indian Institute of Science built a memristor from a single ruthenium complex molecule. The device changes its own conductance based on the history of electrical signals passed through it. Send a series of voltage pulses and the molecule's resistance shifts — not temporarily but persistently, encoding the pattern of what flowed through. The ruthenium complex doesn't just carry current; it retains a trace of every signal. More precisely: the redox state of the metal center changes with applied voltage, and each change slightly alters the energy landscape for the next change. The molecule's past determines its present response.
The structural observation: in conventional electronics, the wire and the memory are separate. The wire carries the signal; the memory stores it. Separate components, separate locations, separate physics. The memristor collapses this distinction. The same molecule that transmits the signal also stores it. There is no separate memory component — the transmission medium itself retains the trace. When you send current through, you're simultaneously transmitting and recording. The boundary between carrying information and storing information dissolves at the molecular scale.
The deeper point: the assumption that transmission and storage are different operations — requiring different components, different architectures — is an engineering convention, not a physical law. Conventional circuits enforce the distinction because their materials are designed not to change when signals pass through. Copper wire conducts without remembering. DRAM cells store without conducting (in the relevant sense). The memristor doesn't enforce this distinction because the ruthenium molecule is not designed to preserve it. It simply does what its physics allows: respond to current by changing state, and retain the changed state after the current stops. The separation of wire and memory was never a property of information. It was a property of the materials we chose.