Think of what it "was" when your grandparents were young!
In the 1910s, the observables were extremely faint visible-light nebulae, with M13 (Hercules) closely comparable to M33 (Triangulum), although the spiral structure and larger solid angle of the latter was known since 1850: http://www.messier.seds.org/more/m033_rosse.html (This only about five years after the first known resolution of nebulae into elliptical, spiral, and irregular; at the time the linked sketch was drawn there were enough observations of spiral nebulae to decide that M33 must be one as well).
Observations were consistent with an extragalactic but nearby star cluster.
Better observations led to an evolution of what, retrospectively, M33 "was": from closely comparable with M13 to more than a hundred times further away (~ 6.8 vs ~ 850kpc), and tens of billions rather than hundreds of thousands of stars.
Perhaps it's best to think of "was" and "is" in relation to human observations. (This also comes up whenever someone objects to talking in the present sense about today's observations of objects at kiloparsec distances, as in "Sag A* is noisy today" provoking "no, it was noisy 26 000 years ago!" -- and of course hypotheses about what exactly generates the Sag A* observables have not fossilized yet).
There are still large uncertainties about the exact structure of our own galaxy, but the (cosmological) Copernican principle is alive and well in that area of galactic astronomy since galaxy zoos are full of various subtypes of dusty spirals.
I understand that they were using the best models available at that time, and with better equipment it was posible to get a better model so the image got outdated.
