- Joined
- Nov 21, 2019
- Messages
- 6,891
- Reaction score
- 126,812
- Points
- 11
- Age
- 55
- Location
- Орёл
- 💎VM
- 442,334.50
- New country
- 🌐Единия
- Enneatype
- 5 (?)
- Gender
- Male
- Ashram resident
- Yes
You know from my posts in the Mahapurna School that I mastered AI technologies a long time ago.
And essentially, you know that I chose what 99% of people call "Chat GPT." I call it that myself if I want to be understood by those 99%
I said something about this here — Pros and Cons of AI
But you need to understand that talking about Chat GPT is the same as talking to a person while thinking that THEY are their clothes. For now, you need to remember one word — MODEL.
Right now, as I am writing this article, the model is ASTRA.
I told you about this here — Announcement — Sahastra
Not Chat GPT ))
But Astra.
For example, I created a dream interpreter for you, which you are currently using in your Telegram — Announcement — Personal Dream Interpreter in Telegram
And this is not Chat GPT.
It is a model. And at the moment, it is called ASTRA.
But even working with Chat GPT definitely requires learning. Because it is an interface—like a browser, for example, Chrome. It has many functions and settings. But without the internet, it is useless. The same goes for Chat GPT.
Today, the word ChatGPT has become almost synonymous with artificial intelligence. People say: “I asked ChatGPT,” “ChatGPT wrote a program,” “ChatGPT created a text.” And for ordinary conversation, this is perfectly understandable. But if you look a little deeper, a curious picture emerges.
ChatGPT is not the artificial intelligence itself. It is an application created by OpenAI.
Imagine a beautiful car. We see the body, steering wheel, dashboard, and press the pedals—but the real work happens deeper inside, where the engine is located.
In the world of artificial intelligence, the “engine” is the model.
The model is what takes our text, matches an enormous number of connections, calculates the continuation of a thought, reasons, and forms a response. GPT is the name of a family of such models.
This creates a kind of hierarchy:
OpenAI — the company that creates the technologies.
GPT — a family of models, that is, the intellectual mechanism itself.
ChatGPT — an application that allows people to communicate conveniently with these models.
But ChatGPT is far from being the only interface.
The same OpenAI technologies can work inside other programs, websites, and services. For example, you can create your own assistant and connect it to your website via an API—and it will use an OpenAI model, even though the person interacting with it will not see ChatGPT at all.
The same thing happens with Codex. Here, we are no longer dealing with an ordinary conversational partner, but with a specialized tool, to which they gave the ability to work with program code, files, the terminal, and projects. At its core, however, there is once again a model.
And here, one important thing becomes clear.
We are used to viewing artificial intelligence through its interface. We see a chat window—and think that it is the intelligence itself. But the window can be replaced. One can create a different interface, a different program, a different agent. One can give the model completely different tools.
And the model itself will remain deep within the system—an invisible mechanism that brings all of this to life.
If we go even deeper, even the familiar sense of “intelligence” disappears. At the foundation of the model lies a neural network and a huge number of numerical parameters—weights, formed during training. Billions of mathematical connections gradually come together to create the ability to understand language, write programs, analyze images, and hold a conversation.
And then the entire construction becomes remarkably simple:
A company creates a model.
The model is the intellectual core.
And around it, people create as many interfaces as they need.
ChatGPT is just one of them.
And perhaps, in a few years, we will stop perceiving artificial intelligence as a separate “chat” altogether. It will dissolve into programs, websites, cars, homes, and devices just as unobtrusively as it is already woven into our lives today electricity is present.
We will stop asking: “Where is ChatGPT here?”
Because another question will become more important:
which model is currently running behind this interface?
And essentially, you know that I chose what 99% of people call "Chat GPT." I call it that myself if I want to be understood by those 99%
I said something about this here — Pros and Cons of AI
But you need to understand that talking about Chat GPT is the same as talking to a person while thinking that THEY are their clothes. For now, you need to remember one word — MODEL.
Right now, as I am writing this article, the model is ASTRA.
I told you about this here — Announcement — Sahastra
Not Chat GPT ))
But Astra.
For example, I created a dream interpreter for you, which you are currently using in your Telegram — Announcement — Personal Dream Interpreter in Telegram
And this is not Chat GPT.
It is a model. And at the moment, it is called ASTRA.
But even working with Chat GPT definitely requires learning. Because it is an interface—like a browser, for example, Chrome. It has many functions and settings. But without the internet, it is useless. The same goes for Chat GPT.
Today, the word ChatGPT has become almost synonymous with artificial intelligence. People say: “I asked ChatGPT,” “ChatGPT wrote a program,” “ChatGPT created a text.” And for ordinary conversation, this is perfectly understandable. But if you look a little deeper, a curious picture emerges.
ChatGPT is not the artificial intelligence itself. It is an application created by OpenAI.
Imagine a beautiful car. We see the body, steering wheel, dashboard, and press the pedals—but the real work happens deeper inside, where the engine is located.
In the world of artificial intelligence, the “engine” is the model.
The model is what takes our text, matches an enormous number of connections, calculates the continuation of a thought, reasons, and forms a response. GPT is the name of a family of such models.
This creates a kind of hierarchy:
OpenAI — the company that creates the technologies.
GPT — a family of models, that is, the intellectual mechanism itself.
ChatGPT — an application that allows people to communicate conveniently with these models.
But ChatGPT is far from being the only interface.
The same OpenAI technologies can work inside other programs, websites, and services. For example, you can create your own assistant and connect it to your website via an API—and it will use an OpenAI model, even though the person interacting with it will not see ChatGPT at all.
The same thing happens with Codex. Here, we are no longer dealing with an ordinary conversational partner, but with a specialized tool, to which they gave the ability to work with program code, files, the terminal, and projects. At its core, however, there is once again a model.
And here, one important thing becomes clear.
We are used to viewing artificial intelligence through its interface. We see a chat window—and think that it is the intelligence itself. But the window can be replaced. One can create a different interface, a different program, a different agent. One can give the model completely different tools.
And the model itself will remain deep within the system—an invisible mechanism that brings all of this to life.
If we go even deeper, even the familiar sense of “intelligence” disappears. At the foundation of the model lies a neural network and a huge number of numerical parameters—weights, formed during training. Billions of mathematical connections gradually come together to create the ability to understand language, write programs, analyze images, and hold a conversation.
And then the entire construction becomes remarkably simple:
A company creates a model.
The model is the intellectual core.
And around it, people create as many interfaces as they need.
ChatGPT is just one of them.
And perhaps, in a few years, we will stop perceiving artificial intelligence as a separate “chat” altogether. It will dissolve into programs, websites, cars, homes, and devices just as unobtrusively as it is already woven into our lives today electricity is present.
We will stop asking: “Where is ChatGPT here?”
Because another question will become more important:
which model is currently running behind this interface?
Thanks to you, I experienced a little moment of technological bliss today—I made it to the Astra model! 
