From 3D GenAI to the End of Game Engines

About this episode

Tejas Kulkarni, CEO of Common Sense Machines and PhD in Brain and Cognitive Sciences and AI from MIT, explains how CSM turns images and sketches into game-engine ready 3D assets. He breaks down why CSM ranked highly on the 3D Arena human evaluation, what “good” means in 3D generation, and how geometry, texture, topology, and parts segmentation affect real workflows. We discuss modular pipelines, UV mapping, controllability, and integration with Blender, Maya, Unity, and Unreal. Tejas contrasts mesh pipelines with splats and NeRFs, then argues that neural network driven game engines are coming, with artists guiding structure while networks fill detail. He also tackles ethics, brain hacking loops, and adoption advice for students and studios. Listeners get concrete guidance on quality metrics, parts based editing, and near term toolchain decisions.

Subscribe to XR AI Spotlight weekly newsletter 👉 https://xraispotlight.substack.com/
***
CONNECT WITH OUR GUEST
⁠💼 ⁠⁠⁠⁠⁠Linkedin⁠⁠⁠⁠⁠⁠
⁠🔗 ⁠⁠⁠⁠⁠Website⁠⁠⁠
 
***
CONNECT WITH ME
🐤⁠⁠⁠⁠⁠⁠X (Twitter)⁠⁠⁠⁠⁠ ⁠
💼 ⁠⁠⁠⁠⁠⁠Linkedin⁠⁠⁠⁠⁠⁠
🔗⁠⁠⁠⁠⁠⁠Link-in Bio⁠⁠⁠⁠⁠⁠
 
***
PAST EPISODES
🎙️ ⁠⁠⁠⁠Show Website⁠⁠⁠⁠
🟢 ⁠⁠⁠⁠⁠⁠YouTube⁠⁠⁠⁠⁠⁠
🍎⁠⁠⁠⁠⁠⁠Apple Podcast