How exactly would you suggest people read the paper? The article linked to a brief [1] which says literally absolutely nothing about how they conjured up their numbers. And the paper itself is here [2]. It's rather worse than a paywall. Instead the paper is restricted to their "sponsors" until an embargo of up to 6 months.
This sort of behavior should not instill confidence in the paper's conclusions. And I think it was completely reckless of the Guardian to run an eye catching headline with absolutely 0 justification given to the numbers other than an implicit appeal to authority by calling it an MIT paper. Though such is the state of the media today.