(EDIT:) The basic difference is that 'we' suggests two minds, specifically a teacher/student paradigm.
This is perfect for any teaching context.
But academic peers exchanging information will go to absurd lengths to avoid insulting their colleagues by insinuating such a relationship.
The object of maths is understanding.
The object of mathematical writing is communicating that understanding with the utmost clarity.
And at the heart of communication is dialogue. It is one person explaining to another person.
Look at the Greek dialogues (e.g. http://www.sacred-texts.com/cla/plato/laches.htm), and observe how easy it is to understand the material when both parties are voiced, even if the student's part is little more than a device.
This duality is critical; question and answer complement one another; the question creates a space into which the answer may manifest, the answer provokes further questions, and so on. This creates an interconnected web of understanding.
And if the 'we' is lost, then the dialogue is also lost.
Look at Einstein's famous papers (e.g. http://milesmathis.com/five.html) and notice how you feel he is looking over your shoulder and explaining something to you.
It is unfortunate that academic papers feel the need to be absolutely impersonal, because it makes them dry. I have to read through a lot of them, and the conventional style is a hindrance; you should break every rule in the pursuit of clarity.
Interestingly, as YouTube videos are gaining popularity over written text as teaching resources, we are seeing a resurgence of the Socratic Dialogue paradigm! As a video is typically made in real-time, the presenter doesn't have the luxury of formalising the presentation. This is a good thing!