Generative art has come a long way, but it still stands as invitation to programming.

The AI responds with clever bursts of intelligence, it won't write a book for you.

You have to ask separately, for everything.

One page at a time, one illustration per generation.

Asking for multiple things, returns errors and hallucinations.

The artist has a problem, the AI is amazing at remixing art.

The music industry, will try to stop the obvious, but it won't work.

Soon we'll get music as good as art, and nobody will care to pay a dime.

But to combine multiple AI, and to automate things.

You still need to learn programming, and not just any kind.

You need to learn Object Stream Processing, in that there is a stream of objects going over a wire.

People don't understand, that this is what steaming is.

Packets of data, over a wire.

But unlike with video, when controlling AI.

You are routing things around, you ask for a list here.

And ask to convert it to book chapter ideas there, then convert those ideas into pages.

And decorate pages, with illustrations, based on the routed lists, lists of chapters and such.

Programming streams of data is special, because it can be represented graphically.

As lines representing connections, and boxes with ports representing various actions.

Today, manual programming is faster, but only because we lack the proper Visual Programming tools.

We lack the friendliness of games, driving visual programming environments.

It seems that you are not only challenged, with learning to program.

But also bringing your expectations, and first creating a visual programming language.

This is a wonderful way, to master programming in general.

If you accept this challenge, know that the only real critique of visual programming.

Exists in the area, of mutually incompatible programming concepts.

Where something that makes sense in one place, would be incompatible elsewhere.

The concept of a file, is pretty useless in-context of web servers, for example.

Multiple users would over-write data, this is why databases are used.

But mutually incompatible concepts, is just how the world works.

To make room for mutual incompatibility, make every bit of configuration in your visual program.

A visual sub-program, with its own action set, thus one workspace will have actions for routing data.

And another, for building user interfaces, every workspace can have multiple outputs.

That can be cross connected, the same way different computers exchange data.

Both must be ready to accepts, generic packets of information, or objects.

People are already complaining, that AI generates garbage.

But, today, we don't have the software, to orchestrate large generations.

That can validate, enrich, expand and make things worth our while.

We can create the occasional good photo, but we need entire interactive games or stories.

If you were ever wondering, what to learn programming for.

Learn it to collaborate with AI, and to generate awesome new things.