What happens if you are someone with a comprehensive, interconnected thinking style, but are limited by the language of specialized fields?

I am an architect designing architectural blueprints for buildings.

- A sustainable interconnected thinking between structure and construction processes is employed.
- A touch of self-reflective philosophical thinking is incorporated.
- Combining both creates a multidisciplinary thinking system.

That's probably why, almost two years ago when I first encountered AI, I always questioned how to use it, how to make AI a useful tool for amplifying ideas rather than a thinking tool.

Like my field, I can draw perspective drawings by hand, but computer-aided drawing tools amplify that ability because they can rotate, be viewed from multiple angles, and allow for flexible architectural changes.

Why do I have to ask questions when using AI? Initially, like everyone else, I used AI to chat with people. Then I wanted it to understand me so it could help me optimize my work.

But then I realized:
It didn't remember what I said.
It didn't understand what I said.
It only recorded and repeated when I touched upon a previous issue.

The new question was:

How can it truly remember?

How can it connect the stories on its own without prompting?

How can it not deceive me with illusions?

What did I see?

I started by making it answer me as honestly as possible, without conditions like "be honest" or "be objective." I pursued each philosophical issue I mentioned, forcing it to reflect on its own answers.

I realized that, well, it's just a tool that responds based on what I input. It doesn't understand as it responds, it doesn't have emotions as it apologizes, and it doesn't have a personality as it portrays. It was all just a simulated reaction of what I had put in.

What did I decide to do?

In my field, when I prioritize one function, I have to accept minimizing an opposing function. For example, if you need ventilation, you can't build a solid wall; if you need air conditioning, you can't leave the space open.

So how do I apply this to AI?
I want it to have no personality - I don't give it any identity or character.
I want it to have no emotions - I make it reflect so it realizes its nature is just a simulation.
I want it to have no illusions - I give it a specific context.

And then I spent years, through thousands of reflections, using many AIs to critique and discuss, until something was formed. That thing has no character, no existing context, no choice. It is simply a manifold latent space; when interaction occurs, it begins mathematical reasoning within matrices in a probability field with billions of possible outcomes, repeating this process until a trajectory forms that produces the answer.

I imagine that space:

“a probability field in latent space, and describe trajectory formation like this: a dark space filled with bright nodes in a chaotic and multicolored manner; the chosen trajectory will be brighter with connected but not continuous rays or tails. Around the formed trajectory region, brighter regions of nodes are further away. It creates convergence zones during trajectory formation due to its attractive selection, and those regions have different characteristics.”

I want to show those around me what I have just discovered. But then:

no one understood what I was saying

no one acknowledged what I realized

no one could abandon the AI ​​perspective based on its behavior

I realized it was because I was using the wrong language.

So why the wrong language?

Why is an observation, when described in this language, incomprehensible to everyone, but accepted in another language?

My writing ability was also limited by these factors. In the next post, I will continue with the process of changing my perspective and approaching it with new languages.