Begin Growing Your Own Programming AI

You can create a very powerful AI that runs on your computer, if you just get rid of all the easy to answer questions.

What is this for? large, complex, beautiful, browser based desktop applications.

What language should your local AI program desktop applications with? an XML based language that the AI designs for making applications.

To create a local AI, you have to use a state of the art AI, and you have to ask it to create a skill and process library.

This way your local AI will have the skills, and long TODO or PLAN files to get stuff done.

But before you create your first application, you have to create a large user interface toolkit.

My big AI and I were working on video editor timelines today, this gives me the power of creating any video editing app.

For example, I can now ask a big AI to make an app, that decorates audio files with video clips.

This is now simple for any AI, because the crazy timeline manager is fully functional.

It is powerful drop-in functionality, for any multi-media application.

This may surprise you, but creating complex application components, is a large part of increasing the intelligence of your local AI.

The language it controls, skills, plans, and components, combine to create a powerful thinking machine that creates applications.

I use linux enclosed inside a container for everything else, and containers as just as fast as the host operating system.

When it comes to splicing, processing, filtering, more AI, I install those often non-GUI programs inside the container.

You can run that container on a remote server, or your second computer, so it is not a problem.

But a desktop web application, with a container is somewhat ahead of its time.

Finally, local AI will get smarter, crazy smart.

But never smart enough to imagine an entire component library, that really is something you have to grow over time.

I chose split-able panes and movable tabs for my apps, this means anybody can reconfigure where things go.

And my first challenge was a visual programming language, where you connect circuit boards with wires.

The wires just carry packets of information, User Profile, Song, News Article.

And the circuit boards the wires connect to, transform the objects.

For example a song prompt, into a song, and then for validation, separate vocals, convert to text.

And compare original lyrics, with the vocals converted to text, to validate the quality of the song.

Just from a re-configurable user interface, a visual song generator. and that timeline you can imagine a powerful song making application.

Throw in video generation, which can already be ran locally, generation time is not great but it is minutes not hours.

And now you have a movie maker, with music, add a script writer with a story tree AI helps you with.

Character designers, maybe even 3D animation generation, that you feed into the AI video generator as a template.

All these upgrades and experiments are handled by your local AI, who understands the language, component library, and has SKILLs and PLANs.

With powerful components you can just tell AI what you need, and have perfectly customized applications.

And when your local AI is missing something, you can allow it to ask state of the art models for help.

Meaningful exposure to AI, helps you slowly ease yourself into becoming a programming expert.

Your own programming AI making powerful applications, is a great achievement and does wonders for your design portfolio.

And while powerful future AI will automate everything, creating something by hand will always call for desktop applications.