I am still learning about game development and development in general. I am reflecting on why I do not like the existing art pipeline in engines. You need to get locked into some decisions. One small example: if you want to do blend meshes, you are locked into the mesh, and you cannot modify the mesh. If you want to make changes, you will need to redo all your blend meshes. These kinds of issues are littered across the pipeline. A tiny change in the early stages of the pipeline causes a ripple effect.
Back in the day, we had two ways to serve JPEG images on the internet. The internet was slow, so you could see how the image was loading. One way was loading from top to bottom, and another way was blurry with little detail and then became sharper over time.
I think in game development (maybe in general software development) it is important to get the rough game from beginning to end, so you will have some blurry vision of the game and a general idea of how the game feels, what kind of weak or strong spots the game has, and in general, if it’s worth continuing to work on the game.
So, the first step is to implement the main game loop and all service functionality around the game as fast as possible with placeholder assets and without any important functionality. This way, you can get the feel of the gameplay. I think the second stage is to add assets to bring the look of the game as close as possible to the final look. This is more to keep excitement about the game fueled by nice aesthetics. I found having a lot of placeholder and temporary art demotivating. Also, once you are done with this step, the game is technically shippable. It is just about adding more content and polishing it up.
As I mentioned at the beginning of the post, I am still learning, and these are just my thoughts, not tested. What are your thoughts?