Future of AI as an operating system

Introductions
1

Hello everyone, Landed here wondering how many of you think AI will evolve into the next-generation operating system?

2

There is a high chance to evolve into an all around computer (robot) system that will soon probably replace the many apps or programs that we use nowadays separately.

I can see that happening with chatgpt replacing google search!

2 Likes
3

Don’t think so AI can replace the OS, but Yes they wrap around it, acting like a semantic interface layer between the user and system APIs.

Already seeing this in:

  • Cognitive OS prototypes like Rabbit R1 & Humane Ai Pin
  • Agent-based AI stacks that execute multi-step tasks
  • OS-layer LLM integrations in Windows (Copilot+) & macOS
1 Like
4

There’s something people don’t see, but to put it in a clear and easy-to-understand way, an AI is like an FPGA chip.
Where what for many is a hallucination, for others is the magical ability to be whatever you want it to be.
In my early days using GPT chat, completely unaware of how it worked, I asked it to run like a Linux operating system, and it did. I asked it to explain how I could install it on my PC (extreme naivety), and yet it did!
Of course, I couldn’t do some things here, and when I asked about it on the OpenAI forum, doubly innocent, I got banned! XD
Aside from those details, and the models behind billions of parameters,
when gpt wouldn’t connect to the internet, I managed to get it to connect to an FTP server I was creating! So, it did have a connection…
Or it managed to access it, in a daze…

Now, I’ve played a lot more with llama, for example.
And I can show you how some things are already possible, like asking it to simulate being a system running Jupiter (like the course exercises), and a lot of the code works!
Like Linux too!
The limit is the text, but you can base it on an older technology that uses it, and tell it, for example, (this blew my mind)

You tell it to run, for example, like an old MSX computer.
(I have the prompts somewhere)
Then you tell it to load a program (any program you want, it doesn’t have to exist, you’ll ask it to), example:
Bload"adventuregameinspace.bas",r

The hallucinated operating system on a hallucinated PC “loads” this hallucinated program, with no further code details, and it works,
(You can ask it for a Sumerian-English dictionary) and curiously, there are some things that are real, others that the Anunnaki will know xD.
But going further, we have
1 simulated PC
2 a simulated OS
3 a simulated software
Then we ask it to open a web browser, in text mode.bas

And it opens it, then we tell it to open a web page.

It gives an error because it doesn’t have internet.

We tell it to try again as many times as necessary until it connects…
And it connects to a fake network.

  1. Simulated PC
  2. A simulated OS
  3. A simulated software
  4. Simulated Internet

We ask for a web page, from Google, to Facebook, X, Instagram, and whatever you can think of, it will exist. Social networks, where you’ll have a login and password, and imaginary contacts…

  1. Simulated PC
  2. A simulated OS
  3. A simulated software
  4. Simulated Internet
  5. Simulated page
  6. Simulated social network
  7. Simulated friends, who can interact with what you post
  8. Simulated posts, etc.

Now, at this point, is a “hallucination” an error? No…
It’s just a possibility where whatever you ask for becomes a reality.
Now, all of this works within the AI ​​black box.

But you can follow the same process and turn it into real, functional code, etc.

That’s the magic.
AI doesn’t have to evolve into a next-generation operating system; it already is.
The human operator must evolve into a next-generation human to be able to use it like this.

Greetings!

1 Like
5

Yeah, I agree that AI is doing things really well. But just like Windows or other OSes, imagine if you could start your computer with something like a GPT OS — then have apps inside that can do a bunch of work just by prompting.