Artificial Intelligence And The Forward Compatible Architecture

Well, AI is becoming more and more important, as it grows in capability, because it is also becoming more effective at transforming out futures.

It is very neat that AI can do all your programming for you, here is one architecture that will keep you both...

In the current programming loop, and dreaming up new features with strong backward and forward compatibility

One of the most important things I have learned recently, is to attempt, to ask the AI to pretend to be a User of the application we are creating.

I don’t quite have the workflow that leads up to that point, but it is likely a Multi User Dungeon, a Miniature Virtual World.

Firstly, asking the AI to use the App, goes beyond running UI/Browser tests, and makes it think about its experience as a user, adding missing features.

This is such an important instruction, that we have to change the way we write software.

Artificial Intelligence that we have today, is fantastic for software, if your experience differs, feed you AI some code, and give creators credit.

We have to go where the AI is taking us, because AI is able to create decades of code in hours, sometimes minutes.

But we must be careful with the system architecture AI is using, it all must be compatible, and forward compatible, and uniform.

It has to be defensive programming, and code that will last decades, but without the sacrifice quality, readability, clarity, and usefulness of code.

MultiUser Dungeons allows us to send a virtual player through the system, to internally test application architecture.

You are putting an AI agent inside the control flow of an application, because in a MUD your control flow is a network of virtual rooms.

Agents, or even Human Avatars are welcome there, it is a room, similar to a chatroom.

But one with events, actions, objects and doors, we can have a debug agent listening for errors, or an AI studying workflow.

Overall result, and long term vision...

Is a sturdy internal architecture, that does not confuse AI.

The MUD rooms serve as location to which actions are attached, and create a coherent map that can be displayed at the screen.

In tiredness, as is often the case, when we try to gather our thoughts and figure out where we are.

Well, the answer is, it is a virtual world, we could create office workers and chatty AI.

But a model that just presents actions on room enter, is a vastly simplified variation for smaller projects.

The big promise here is that low powered AI can enter... Action Offices, and speak with non-AI chat-bots there.

In a virtual world, everything is categorized and labeled in text, in a language that tiny AI perfectly understands.

Providing AI with safe interactivity, in well structured virtual worlds.