It would be great to have a generation type dialed to 3D printing - this doesn't need a texture at all but you could use the texture data to make a height map and apply it to the mesh to add an incredible amount of detail. On the image here I show how I manually used the texture map of a prompted lizard warrior to enhance the mesh itself. This is just a filter that could be algorithmically built in but I believe the AI itself could achieve even better fidelity
I also noticed that using images of parts of a character (hand/head) give much better detail than a full model. It would be great if it was possible to enhance the details of one generation, treating it like 4 or more (similarly to 2D generators that allow you to enhance resolution etc)
Finally I noticed the image to 3D generation tends to break if it's a drawing that's too flat. Using another 2D AI allowed me to first edit the sketch to look like a render, which generated fine in 3D. If Tripo could catch if the mesh generated is too flat and try to morph the image to create depth it would be an incredible boost to its efficiency.