They edit: ~~aren't~~ are barely domesticated. They are useful already in their wild form. They can be "broken" to recognize a dominate rider as their herd leader, but they have to always be tied or corralled. People are trained on how to approach horses, not vice-versa. Never approach a horse from behind.
Domesticating seems like a fuzzy concept here, in the way it's being discussed. For, as dog owners, we too (well, the more.. principled ones) are trained in how to approach a pup -- or a 'misbehaving' adult dog -- and how to train/discipline the dog in turn.
As someone who has owned a pregnant mare, and observed (and been involved where appropriate) in the birthing, raising, and, with time, subsequent breaking of the foal, I think horses may be born in a more 'wild' form--- but only in the loose, as-it-were sense of the word.
This all said, domestication should probably be discussed in a way that's wholly specific to a species since, with the horse-dog comparison, it goes without saying that their cognitive development, experience, functional relation to humans, and genetics (unlike the wolf/dog in this regard, a premise of the article) differ in such profound ways.
Yes, but I mean to a far lesser degree than a dog. You get utility very quickly with a horse for very little trait selection. For example, you can get this far (no traits selected, completely wild, not feral) with a never-domesticated Przewalski's https://www.youtube.com/watch?v=ovUGagVGiZ4