I’m certain we’re missing something in that article. An electric bike shouldn’t be more efficient than a normal bike - it’s heavier and the wind resistance is the same. My first guess is that the power use for the electric assist only considers the electric assist, not the human riders portion.

Humans are quite inefficient - we can convert around 25% of food energy to work. The losses in the power plant generating the electricity for charging the e-bike's battery and the motor's efficiency losses would result in a higher overall efficiency than 25%.

People generally dont need more calories that they already have, they need less. They must do (but dont do enough) physical exercice to loose extra energy they consume (in places where cars are an issue).

Sure, but that doesn't really play into the equation when comparing the pure efficiency of different transportation methods. Of course in practice since it's good to exercise at least 30min daily, you can think of the energy spent during that as free from a efficiency standpoint

Heavier would seem like a benefit to me, wouldn't inertia improve its score over longer distances?

No, the only things that matter are the losses and those are due to air resistance (same), rolling resistance (increases linear with weight) and losses due to braking/acceleration (grows quadratic with weight). Heavier is always worse. Note that electric bikes usually don’t even have recuperative braking due to the complexities involved.

If heavier were better, we’d see all Tour de France riders use cast iron wheels and frames ;)

You might be thinking of heavier race cars with bigger engines generally having higher top speeds. They aren't more efficient, just more powerful, and weight doesn't affect top speed (only acceleration).