The general AI can be a networked cloud-based system.

 The general AI can be a networked cloud-based system. 

The general AI can be a cloud cloud-based system that connects sub-systems under one umbrella. Those sub-systems can be independently operating AIs that central AI can connect to work with the same solution. 

Every data system involves hardware and software. In cloud-based AI every part of that system can close the problem from different angles depending on what type of datasets each system can use. 

The problem with monolithic architecture is if the server falls the entire system stops its work. Web-based architecture with multiple cores guarantees that the AI works even if there is some kind of problem in one server. 

The cloud-based architecture makes the networked AI an ultimate power.  We can think that the future belongs to limited and precise AIs. The fact is that the AI-based chatbots can also interconnect those independently operating AIs under one umbrella. A limited AI could be the AI-based chatbots that collect data for some special missions, like stock marketing and financial investments. 

The thing that determines what purpose the AI has is the dataset that the AI uses. If the AI must only follow stock marketing homepages and compile numbers from those pages. That thing makes it very accurate. But it's possible to connect that limited AI with some other AI that can use open sources for searching data. 

That open source can be things like the Financial Times and some news pages. The AI can search for information from the net about those companies on the stock marketing pages. The reason for that is, big investors might want to see what their investment or "investment that they want" makes. If a company is on some kind of boycott list, that can cause problems that take a person into the courtroom.  

The open internet is problematic for the AI because it should select trusted sources. However, the human operators or creators can determine what pages the AI must use.

However, cloud-based AIs can be limited to AIs that are operating under one AI that acts like an umbrella that interconnects their abilities. The idea is like in all other cloud-based structures. 

The cloud-based AI can be networked servers, and each of those servers runs its own AI software. People can see network benefits from that thing. The AI software is a complicated tool. And in a networked solution where each server runs independent AI, there is no need for maximum capable servers. Each server shares responsibilities and resources over the network and that means those multi-core systems make the individual server's work lighter. 

The general AI can be networked precisely created limited AIs. In that version, the AI that we see from outside is an entirety of multiple networked AIs. This means that the networked AI can look like a cloud-based solution. 

From the outside, it looks like a monolithic structure, but that entirety is multiple independently operating AIs that can act as an entirety, or they can act as separated cells that can operate with different independent operations. This makes the cloud-type networked AI an interesting tool. Every part of that network can installed on their servers. In that case, the AI is the server network. Each of those servers runs an independent AI program, as I wrote before.