Since the unveiling of 'GPTs' at OpenAI's first developer event last week, creating custom chatbots has become a trend. With the Assistant API feature, it's now possible to implement knowledge search without the need for document embedding and search algorithms, and thanks to the increased input token length, very long files can now be uploaded in their entirety.

This newly updated ChatGPT feature can replace many existing services of startups, and currently, in Silicon Valley, there's a meme-like saying, "OpenAI killed my startup." So, what services is OpenAI preparing after creating 'custom chatbots'? Today, let's explore the concept of a Large Language Model as an Operating System, 'LLM OS.'

'The horizontal illustration of a Command Line Interface (CLI) display example is ready, featuring a wide vintage computer monitor with a text-based interface'

Image. DALL E·3

Computer operating systems and interfaces have been evolving steadily over the past decades. Initially, Command-Line Interfaces (CLI) were widely used. Users had to enter precise text commands to use computers, a method that required a high level of technical knowledge. Examples of CLI include DOS, Command Prompt, and the Terminal in Windows and Mac OS. CLI was efficient and powerful but not intuitive, limiting early computer use to experts.

Then came the Graphical User Interface (GUI). GUIs allowed for visual interactions through icons, windows, and menus. With GUIs, anyone could open files and run applications with a mouse click, greatly contributing to the personal computer's distribution and popularization.

Recently, Large Language Models (LLM) are gaining attention as the next generation operating system. An AI-based operating system that understands and generates natural language can autonomously perform tasks based on user questions and commands. This provides a richer and more personalized user experience and can learn users' preferences and styles over time.