Generative AI tools like TRELLIS can produce creative 3D designs but often lack physical feasibility, leading to unusable objects when fabricated. To address this, MIT CSAIL introduced PhysiOpt, which augments generative models with physics simulations to verify and refine designs for stability and usability.
PhysiOpt processes input prompts or images to generate 3D models optimized through small structural adjustments that maintain appearance while ensuring durability. This system enables rapid creation of functional personal items, such as 3D-printed cups or keyholders, improving AI-generated object viability for everyday use.
Source & further reading
Mixing generative AI with physics to create personal items that work in the real world ↗Source published 2026-02-25. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.