Those folks know enough to leverage AI as a feedback engine. Either we strive for the same foundational fundamentals, or leap into AI believing our thoughts, once accelerated, yield better fruits.
Is it better to code the 2d game engine from scratch, or brush all that away because the game itself was more important?
Or, is the gameplay informed by the connection between the contraints and features of the engine; does or should it matter to the game designer how many sprites are allowed, or that the tooling deploys every Friday?
If the game matters, delivery matters.
If the journey matters, and the game was just the medium for applying software ideas, then the craft matters.