This argument is totally valid, but I think the problem is that it's universal. It seems like it's equally valid as an argument against any feature. So to really apply it in practice, we need to go further, and come up with some way to quantify the value a new feature brings, so that we can compare that value to the downside of more bloat.