One of the areas we’ve been playing with is “text to CAD” or “text to surface” with tools like Meshy or AdamCAD. Working with teams like Cove Architectures, gives a sense for why this might be hard.
It’s one thing to get a great looking rendering, but it’s something else to get accurate surfaces and parametric models. The image above is a few steps into trying to get a 3D printed design from ChatGPT.
The current effort is shown below. a piece of aluminum from what I think was a tiling edge/guide. Cut and bent to fit the original pivot mounts for the popup lights. then the 2x projector LED lamps are just bolted to this with their original mounts. Some hose clamps are bent and epoxied to the metal to support the heat-gun-bent plexy cover.
What I’d like to do is print a single piece that can also include a brake cooling intake duct.
How to best describe and give feedback? Seems like a sketch might help versus only text. Or an image but with some ability to circle and highlight the features to which the mount should attach? Will keep playing but relative to other prompts this seems MUUUUCH more difficult to get even an ok output.
Couple more efforts below.
Outrun Carbon 