The use of AI is clearly underrated in a few aspects by a lot of people because they think AI cannot reason as clearly as they can about some of the problems they have in life.

But what if AI is not exactly a tool for solving your problems?

What if it is more useful as a tool for finding the right problem to solve in the situation you are in?

I’ve been seeing concerns about people getting trapped in distorted perceptions through AI because these systems can be too sycophantic toward the user, creating a kind of bubble of perception. Okay, that’s a huge problem. But when you are conscious of this possibility, I think a lot of things change.

When you know that a tool can be designed to please you, you can also think about how to use it to criticize you.

I see AI as a tool to help me understand my experience in the world better. I see it as a tool to help me understand how people’s mental models are created and how I can build a better mental model based on reality, even though my own experiences inevitably influence it with an intensity that is hard to define.

So I started asking myself a different kind of question.

What if my perception of the world is not actually correlated with the real world?

What if social media shaped my mind in a certain way? By the way, I’m 23 years old. I grew up during the rise of social media. What if that affected my capacity to experience reality in ways I don’t fully understand?

And what if I can use AI—with access to the internet and good scientific articles—to help me shape my mind into something better for myself?

Of course, there is a difference between receiving information by listening, reading, or watching something and learning through experience itself.

So what if I use AI to experience more?

What if I use AI to help me approach emotional problems and understand their effects on my daily life?

Just to be clear: I’m not saying that AI is a therapist or psychiatrist, or even that it should be. But I think these professionals, with good use of AI, may be able to reach the root of certain problems faster than we imagine.

And then I started seeing a pattern in all of these questions. I think you probably saw it too.

I’m using AI to solve problems that I think will help me.

But are these the real problems I should be solving?

Is this the right square?

Is this the variable that, if changed, will actually have the greatest influence on the equation of my life?

So what if I change the approach?

Instead of going to Claude, ChatGPT, Grok, Gemini, or another AI and saying:

“I have this problem. Help me solve it.”

What if I give it the context and say:

“I’ve been trying to solve this problem, and I’ve created a lot of solutions based on my assumption that this is the problem. But what if the actual problem is not one of them? Could you help me figure that out based on the context I’m giving you?”

Now AI is doing something different.

It is not only generating solutions inside the square you selected.

It is helping you ask whether you selected the right square in the first place.

And then, if you’re lucky enough, the probability of the next word might have a much bigger impact on your life than you expected.

To publish a new post, I just add a Markdown file to _posts/ — nothing else to do.