Does anyone actually make these arguments? I can easily see someone arguing that if you trust your developers then making them make pull requests instead of just allowing them to push changes is an unnecessary extra step (or two). But how could anyone argue that a tool designed for a low-trust environment must be unsuitable for a high trust environment simply because of its origin? Couldn't such an argument be refuted by saying "that's just stupid"?

And who says the web should not be used for apps?