Amazed by complexity

I took this first photo casually, and thinking about it later, found it funny and could have added the ashtray and the revolver.
I was surprised by the quality of the image insertion, but I couldn't understand why the final result gives the feeling that the waffle and the knife looks fake and the ashtray with the cigarette looks so realistic. Would this nuance be due to the amount of online photos used as a basis for the algorithm inducing a more "real" approximation?