Personally, I found the 1955 Gould performance jarring on my nerves. It's very sharp-edged; hard to imagine how it fits with Bach's purpose in writing the Variations in the first place.
Not that it's terrible. It's a virtuoso performance, to be sure; athletic and emphatic.
Dinnerstein's new Variations, to judge from the excerpts (I haven't heard the rest), is by contrast much gentler with the material. I would go so far as to say that a warm Romantic sensibility is present - which may account for the observations you and the New York Times made about certain passages resembling the works of later composers.
But that Romantic tinge didn't come from a prescient Bach. His music has more in common with mathematics and machine language than with the emotion-laden, religion-inspired music which mostly preceded him, or the humanistic Romantic era music which followed. He was the Copernicus of music, covertly worshiping at the altar of mathematics and science.
Who has it right?
I don't think anyone has it right. Of the two, Gould probably comes closer to the historical Bach, as his phrasing is not pregnant with emotion. But you're right, Gould's performance is a 3-alarm fire, not a lullaby. He missed the target.
But Dinnerstein missed, too. A Bach lullaby is not the same as a lullaby from the Romantic era. It wasn't emotion Bach was striving for, but a kind of auditory mathematical space within which the mind could achieve calm. He was definitely not playing for a baby in his belly. His purpose was more akin to hypnosis than gushing saccharine into our ears.
Thus my opinion is mixed. Dinnerstein's phrasing is a guilty pleasure. It's very pleasant, I think more so than in Gould's justly famous recording. That by itself is an accomplishment of note; anyone who can come away with listeners as happy as Gould's is doing darned good. But it isn't the ultimate Goldberg. That mountain has not been scaled.