I think they are doing that because using real images the model changes the face. So that problem is removed if the initial image doesn't show the face
This genereation seems that is getting performance using more power and more cores. Not really an architectural change but only packing more things in the chip that require more power.
Gaussian splatting transform images to a cloud points. GPUs can render these points but it is a very slow process. You need to transform the cloud points to meshes. So basically is the initial process to capture environments before converting them to 3D meshes that the GPUs can use for anything you want. It is much cheaper to use pictures to have a 3D representantion of an object or environment than buying professional stuff.
Basically you train a model per each set of images. The model is a neural network able to render the final image. Different images will require different trained models. Initial gaussian splatting models took hours to train, last year models took minutes to train. I am not sure how much this one takes, but it should be between minutes and hours (and probably more close to minutes than hours).
Board games have been used in AI since the beginning. They provide a good environment as we know the rules and control them. Also as everybody uses them it is easier to compare different algorithms. Most advances in AI were done in board games. Most probably chatGPT uses lot of those things you think are irrelevant in board games (reinforcement learning for fine tuning the responses with human feedback, same algorithms used in board games).