Research proposes the concept of neural computers, exploring emerging machine forms that unify computation, memory, and I/O

ME News Report, April 10 (UTC+8), recently, a study proposed an emerging concept called “Neural Computer” (NCs), which is a new machine form that unifies computation, memory, and input/output within learned runtime states. Its long-term goal is to achieve a “Fully Neural Computer” (CNC) capable of stable execution, explicit reprogramming, and persistent capability reuse. As an initial exploration, the study instantiated neural computers as video models, generating screen frame sequences based on commands, pixels, and user interactions within command-line and graphical user interface setups. The results show that the learned runtime can acquire early interface primitives, especially input/output alignment and short-term control, but issues such as regular reuse, controlled updates, and symbol stability still need to be addressed. The article outlines a development roadmap around these challenges and suggests that if these challenges are overcome, a fully neural computer could establish a new computing paradigm that surpasses current intelligent agents, world models, and traditional computers. (Source: InFoQ)