LegoGPT, developed by Carnegie Mellon University researchers, translates text prompts into physically stable Lego designs. While it operates with a limited set of eight brick types within a 20x20x20 grid, it effectively creates models that can be built. The process begins with generating a 3D mesh from the user's prompt, followed by voxelization and structural optimization to ensure stability. Researchers envision a future where such generative technologies can create personalized products based on user input, while addressing the challenges of real-world assembly and stability.
"We augment each shape with multiple structural variations by randomizing the brick layout while preserving the overall shape," the team explained.
Researcher Ava Pun told The Register: "We're dreaming of a future where making stuff becomes super personal! Imagine you just type what you want or show us a picture of a chair, and boom - we could actually make that product and ship it to you in just a week or two."
Collection
[
|
...
]