OpenAI releases Point-E, an AI that generates 3D models

well well well… well, OpenAI is like Steph Curry they can’t stop shooting 3 pointers or should I say 3D point E’s!

OpenAI just released a new AI, that generates freaking 3D Models. Essentially, Point-E is a machine learning system that can generate 3D models in 1 to 2 minutes on a single Nvidia V100 GPU.

don’t wanna bore you on how it works buuuuuutttttt… Point-E doesn’t create 3D objects in the traditional sense. Rather, it generates point clouds or discrete sets of data points in space that represent a 3D shape. So like instead of the typical way 3D images are generated which takes a ton of time to render, Point E does it a lot more E-fficiently (The “E” in Point-E is short for “efficiency,” because it’s seemingly faster than the previous 3D object generation approaches.)

So instead of rendering, point E creates point clouds that are easier to synthesize for computers, the limitation to this is they don’t capture the true 3D essence of whatever is generated.

To get around this limitation, the Point-E team trained an additional AI system to convert Point-E’s point clouds to meshes. (Meshes — the collections of vertices, edges, and faces that define an object — are commonly used in 3D modeling and design.) But they note in the paper that the model can sometimes miss certain parts of objects, resulting in blocky or distorted shapes.

This is from the AI With Vibes Newsletter, read the full issue here:
https://aiwithvibes.beehiiv.com/p/openai-releases-new-3d-generator-ai-study-shows-ai-helps-devs-produce-buggy-code