Goodbye Windows, Hello LLMs: The Future of Operating Systems
Large Language Models (LLMs) such as ChatGPT have fundamentally changed the way we interact with computers. However, I believe that many experts are underestimating the true magnitude of this shift. My bold prediction is that LLMs will replace the core components of our computing experience — the operating system, App Store, and even apps will be replaced by LLMs.
LLMs excel in understanding and manipulating language and data. An LLM-based operating system can dynamically respond to natural language, rendering traditional apps and the App Store obsolete. LLMs can create personalized applications on-demand, removing installation needs, and effortlessly manage various data formats, making specific apps redundant. Join me as we explore this paradigm shift and the limitless possibilities ahead.
The Evolution of Human-Computer Interaction: from command line to conversational interfaces
Human-computer interaction has followed a distinct pattern throughout history. Initially, users had to adapt to the expectations of computers, struggling with systems like punch cards and command-line interfaces that required precise instructions. The advent of GUIs (Graphical User Interfaces) marked a significant milestone, making computers more accessible to non-technical users.
Nonetheless, GUIs have only somewhat narrowed the divide between humans and computers, as users are required to navigate menus and app structures, essentially following the developer’s predefined logic. We’ve all experienced moments of frustration while searching for a specific button or menu to perform a simple action. GUIs require users to think like developers, limiting the user experience and imposing certain behavior patterns.
With the introduction of LLMs, we have reached a stage where computers can fully adapt to users. These conversational interfaces require no behavior change from users, allowing for a truly intuitive and personalized experience.
Chatbots, whether standalone AI assistants or integrated features within applications, have become powerful tools in the realm of AI. Standalone chatbots efficiently handle tasks and interactions, providing prompt responses and ensuring smooth user experiences. Integration of chatbots into existing applications automates tasks, enhancing capabilities and simplifying user interactions.
However, this is just the first generation of LLM applications. We have only scratched the surface of what LLMs can achieve.
Windows Copilot
Microsoft recently announced Windows Copilot. Copilot brings centralized AI assistance to users, allowing them to focus on their tasks without the hassle of switching between applications. With asy access from the taskbar, Windows Copilot acts as a personal assistant, providing seamless access to familiar Windows features and enabling new capabilities like content rewriting and summarization.
While the initial version of Copilot functions as a powerful chatbot, it represents just a fraction of its vast potential capabilities.
The Disappearing Line between OS and Apps
A major result of integrating LLMs is the blurring of lines between operating systems and traditional apps. ChatGPT’s capabilities already go beyond simple text generation. For instance, it can create custom, bespoke apps based on users’ needs.
Here are three examples that already work today:
The current landscape of applications generated by LLMs may seem relatively simple, but the rapid pace of developments in this field promises a future of increasingly sophisticated possibilities.
The great (software) filter
But how does this transformation work? Let’s take a closer look at how traditional apps function. Apps manipulate data and present it through a user interface, requiring users to rely on specific applications for different file formats. This is where LLMs demonstrate their prowess. With their exceptional language understanding and data manipulation capabilities, LLMs effortlessly read file formats and comprehend how to update the data they contain.
Imagine a future where LLMs power the operating system. In this scenario, you won’t need separate applications like Microsoft Word or Google Sheets to edit documents or spreadsheets. The operating system itself will possess the native ability to handle these cases. Simply provide the document or API, and the LLM-powered OS will intuitively manipulate the data according to your needs.
LLMs will handle any assigned task, even those not covered by existing apps, eliminating the need for traditional apps altogether. This seamless integration of data manipulation and user interface capabilities within the operating system heralds a new era of efficiency and flexibility in our digital interactions.
This revolutionary step transcends the idea that LLMs will assist with programming apps; instead, they will have the capability to generate any bespoke app on the fly, rendering both apps and app stores redundant. LLMs as an operating system will serve as the great software filter, causing a mass extinction event in software. The ability to generate custom applications tailored to each user’s specific requirements opens up a world of endless possibilities.
Get a sneak peak of the future: Join our waitlist
The emergence of LLMs marks a pivotal moment in human-computer interaction. With their extraordinary language understanding and manipulation capabilities, LLMs have the potential to reshape our digital existence. As we explore the limitless possibilities of LLMs, it becomes increasingly clear that the distinction between operating systems, apps, and the App Store will vanish.
However, there are still significant hurdles to overcome. Local execution for privacy, memory and performance optimization, and the development of a local vector database for long-term memory are among the key components that need to be addressed. By tackling these challenges head-on, we can pave the way for a future where LLMs seamlessly integrate into our operating systems, revolutionizing the way we interact with computers and unlocking a new era of personalized and intelligent digital experiences.
We invite you to share your opinions on whether LLMs will replace operating systems and who will dominate this new market — established giants or innovative startups.
Would you like to explore the future of operating systems firsthand? Complete this form to join the waitlist and follow me on Twitter.
Thank you for proofreading: Javi Lopez, Benjamin de Kraker, Arjan Haring
