The current logical machinery is far from perfection.

48.32% of the global population never used Google;

45% of the world's population do not own a personal laptop;

39% of people globally do not use smartphones;

33.8% of human beings have no access to the internet.

A new type of machine is needed,

to empower all people in this world,

to give them opportunity to experience the beauty and the value of it.

It must be

• publicly available with no cost (open sourced from ground up)
• locally runnable with no need for access to internet
• intuitive to use and relying solely on human instinct
• responsive & helpful in pure information aspects or physical aspects

With the advancement in probabilistic neural networks, we argue that now is the time to investigate whether it is possible for machinery to be more universally accessible, valuable, and friendly, equipped with purely neural network-centric components from hardware to software, including chips, operating systems, applications, etc.

This machine has its core with only neural networks running on it to drive any input and output processing, which is probabilistic in its essence. It will process, store, and compute multimodal input in neural network weights or other ways vastly different from logical computing machinery (LCMs quoted from Alan Turing), serving as a multimodal comprehension and recomposing machine.