tevfik's notion


OpenAI just realised their latest model, Point-E, for creating 3d complex clouds from prompts.


I played with the model over HuggingFace. The model is still rudimentary, but promising for the future.

My takeaways from this:

  1. We are moving away from static 2d image generation to a 3d world.
  2. 3d models will help us to create better AI-generated video during the changing position of view.
  3. 3d models will open us new doors beyond regular interfaces (laptops, tablets, phones) to especially more exciting interfaces (glasses) via AR and VR.