4 ms·

I wonder if, in a decade or two, we'll witness the emergence of a new OS built entirely upon LLMs from the ground up. Today's Gen-Z predominantly enjoy Chromebo

I wonder if, in a decade or two, we'll witness the emergence of a new OS built entirely upon LLMs from the ground up. Today's Gen-Z predominantly enjoy Chromebooks, smartphones, and iPads, all of which offer heavy abstraction in comparison to traditional desktop OS environments. Future generations are likely to exhibit significantly reduced inclination towards manual interaction with individual apps, folders, and settings. This shift could resemble the philosophical distinction between macOS and iOS today, but amplified exponentially. At that point, existing platforms such as Windows, iOS, macOS, and Android may face obsolescence, sharing a fate akin to that of Windows Mobile, Symbian OS, and BlackBerry OS following the advent of the iPhone—unless one of Microsoft, Google, or Apple venture into this uncharted territory.

dboreham 3y ago
I can type quicker than I can talk and I can read quicker than I can listen to someone talking. But it would be nice if MS could fix their 30 year old taskbar auto-hide bugs in the next 20 years..
carlosjobim 3y ago
Traditional operating systems are extremely abstracted as well. To book a flight online right now, you click a series of coloured rectangles on your screen with a cursor that you control by dragging a mouse about on your desk, or rubbing your finger on a piece of glass. Compare this to what would happen in real life, which is you talked to a person and they would write your name on a flight list.
theamk 3y ago
I don't think so? The files and folders are still there on Chromebooks and Android smartphones, and there is even a stock file manager app, but I find it very rarely used. And individual apps have been dying for many years already with everything moving to the web - for example I have only 3 apps I use daily (web browser, mail client, terminal), and many people only use the browser. This leaves only general "browsing", that could be described as: - There is a virtual document which can be big or small - This document may consist of text, images or videos - At any given time, a portion of the document is displayed on the screen. Navigation commands (keyboard, mouse, special hardware, gestures, etc...) could be used to choose which portion of the document is shown. - Some parts of document can be interacted with (using mouse, or light pen or fingers): then some action happens or a different document is shown This model has been unchanged for at least 50 years - first Smalltalks worked like that, and today's browsers work the same. Sure, there are more colors and images are much higher quality, but the essential interactions are the same. So I am not sure what would exactly LLMs bring to that picture.. sure the search bar may get replaced by a natural-language query, but that won't need the new OS, it's just an incremental improvement of existing status quo.

dboreham 3y ago

I can type quicker than I can talk and I can read quicker than I can listen to someone talking. But it would be nice if MS could fix their 30 year old taskbar auto-hide bugs in the next 20 years..

carlosjobim 3y ago

Traditional operating systems are extremely abstracted as well. To book a flight online right now, you click a series of coloured rectangles on your screen with a cursor that you control by dragging a mouse about on your desk, or rubbing your finger on a piece of glass. Compare this to what would happen in real life, which is you talked to a person and they would write your name on a flight list.

theamk 3y ago

I don't think so? The files and folders are still there on Chromebooks and Android smartphones, and there is even a stock file manager app, but I find it very rarely used. And individual apps have been dying for many years already with everything moving to the web - for example I have only 3 apps I use daily (web browser, mail client, terminal), and many people only use the browser. This leaves only general "browsing", that could be described as: - There is a virtual document which can be big or small - This document may consist of text, images or videos - At any given time, a portion of the document is displayed on the screen. Navigation commands (keyboard, mouse, special hardware, gestures, etc...) could be used to choose which portion of the document is shown. - Some parts of document can be interacted with (using mouse, or light pen or fingers): then some action happens or a different document is shown This model has been unchanged for at least 50 years - first Smalltalks worked like that, and today's browsers work the same. Sure, there are more colors and images are much higher quality, but the essential interactions are the same. So I am not sure what would exactly LLMs bring to that picture.. sure the search bar may get replaced by a natural-language query, but that won't need the new OS, it's just an incremental improvement of existing status quo.