Kosmik, the AI workspace



AI Summary

  • Introduction: Presentation of prototypes for Cosmic, aiming to evolve it into a proactive canvas. The plan is divided into three steps: 1) Creating a multimedia canvas, 2) Making it multiplayer, and 3) Developing proactive features to anticipate user needs.

  • Vision for Cosmic: The goal is to merge a canvas with a language model (LLM) into a fluid workspace that could replace traditional software categories like browsers and PDF readers.

  • Demonstration of Prototypes:

    • Users can import images into Cosmic for mood boards, which can then be organized automatically by the system.
    • The system allows for commands to be issued to organize content dynamically based on user input.
    • A video feature is showcased that transcribes audio to text, recognizing speakers like Steve Jobs.
  • Functionality Highlights:

    • Users can summarize PDFs by right-clicking and selecting a summarization command.
    • Users can run comparative analyses between different files and extract snippets for further exploration.
    • Recommended searches can be generated based on content, facilitating research.
  • Future Plans: Plans to implement features where users can drag and drop elements into the canvas, creating a nonlinear and fluid research process.

  • Conclusion: Excitement for the upcoming batch of AI functionalities (AI stickers) in Cosmic, with an expected release date in mid-March. Encouragement for viewer feedback on the prototypes and emphasis on rethinking the research process in the context of AI tools.