"sub" meshes vs one monolithic mesh - Which technique should I use?

So I've done 3D modelling in 3DS Max and I'm currently learning blender, and I'm beginning to look at creating more detailed models and learning about stuff such as managing topology.

https://www.youtube.com/watch?v=rEj1uHSu1Bw

I've noticed in videos such as this, instead of merging the meshes together and managing the topology of each detail/element, the modeller just creates a new mesh. Is this right? Or is there some benefit to combining these details into one "monolithic" mesh and then fixing any topology issues.

Is there much performance or other downside to creating multiple separate meshes like in the video? Am I overthinking this?