Autodesk has integrated generative artificial intelligence into its Flow Studio platform. The new feature, Wonder 3D, allows creating simple 3D models from text or an image, as well as optimizing and texturizing existing assets. For video game developers, this presents itself as an ideation and quick blocking tool, whose results are exported directly to Blender, Maya, or Unreal Engine for final refinement. It is not a complete solution, but it is an accelerator for the early stages of production.
Integration into the Asset Development Pipeline 🛠️
Wonder 3D operates within Flow Studio, a cloud suite for integrating video characters. Its true value for game devs lies in exporting to standard software. You can generate a basic 3D concept from a textual description and take it immediately to Blender to retopologize, sculpt details, and prepare the UVs. For indies or small teams, this dramatically speeds up the visual prototyping phase, creating design blocks to iterate on. However, the models have low-medium detail and require extensive manual work to achieve game quality, functioning as a base, not as a final product.
An Accelerator, Not a Replacement ⚡
The tool confirms a trend: AI as an assistant in tedious initial tasks. Wonder 3D is ideal for generating variations of an asset, testing silhouettes, or creating base textures, freeing up time for high-end art. Its credit model and opacity about its training data are points to consider. In conclusion, it is a practical addition for preproduction and teams with limited resources, but the artist's judgment and technique remain irreplaceable for polishing any asset to the standard required in a professional video game.
How can Autodesk's Wonder 3D revolutionize the 3D asset prototyping workflow for independent video game developers? 🤔
(P.S.: a game developer is someone who spends 1000 hours making a game that people complete in 2)