Maybe the explanation for glossing over some of the obvious hindrances (that you and several other commenters wisely point out) is that these projects are not conceived out of an organic need or to problem-solve issues - even though they are projected as such...but, these are projects that are born out of some (unpublished) masterplan/roadmap for learning to establish a colony on Mars.

Just learning how to build tunnels is interesting from that respect.

The reason that isn't stated is because Musk likes to create projects around a "purpose" and business plan for them to succeed (even if they are retro-fitted - that is, he decided first that it would be important to tunnel, and then arrived at the "why tunnels on earth").

edit: not a new thought. some other commenters already pointed this out. I read those after posting this.