If they aren't stable, then natural ones would not last long enough for us to ever observe them, except in terms of an energy burst across teh cosmos when they collapse, so that's a bit moot.
Well, true, but keep in mind that the "artificial" ones we're talking about (like the metric Kip Thorne and collaborators used for visualization in
Interstellar) aren't stable either. I do believe there are stable ones, but they're not actually easy to find in the literature. My guess is they're complicated and require more or weirder "exotic matter" to make. And we also have to remember that this exotic matter may not exist. And that Morris & Thorne proved that at least a large class of wormholes are automatically time machines and therefore may be impossible. It's conjectured after all that quantum gravity obeys a form of temporal cosmic censorship.
Well, in that case, these aren't really wormholes in the sense most people expect. This is effectively, "travel to a different universe," not ,"travel a distance in your own universe". I then becomes impossible to have a long route through normal space, and a short one through a wormhole. But, I've heard good folks more recent than Thorne speak about such... so I think there's something missing here.
True. The rigorous but general definition of a wormhole seems to be something like this: if you consider spherical coordinates in flat space, concentric spheres around the origin shrink in surface area as you approach that origin. In a wormhole space(time), those concentric spheres first shrink and then expand as you move "toward the origin." The expanding region is when you've gone "past the origin" and into the "second universe." But with this definition, there doesn't even have to be a full second universe. The whole thing could just be a funny-shaped dimple on a single spacetime.
But keep in mind that I'm talking about just the mathematical formulae people write down because they're simple enough to give exactly. People draw pictures that basically fold over "universe 2" and glue it to "universe 1," which gives you a tunnel between two points on the same space. That's the more colloquial expectation for a wormhole. I just don't think it's feasible to write a formula for a spacetime like that. On the other hand, I don't see a reason that a mathematical description of such a space can't exist --- it'd just be very ugly to write down and not as enlightening as these "toy wormholes."
And it that raises a whole lot of questions - like how a distinct and separate space just happens to have physical laws compatible enough with those in our space to support the connection? And how would this be different from a bottleneck region in our own universe?
Well, even the "two-universe" wormholes are really a single universe/spacetime as long as the wormhole is there. It holds the whole spacetime together. Another simplification of all the wormhole spacetimes I've found written down today is that they are time-independent, meaning they sit there forever unchanging. As for physical laws being compatible, I suppose that depends what you mean. But it's a premature question in some ways, since even these toy models rely either on exotic matter that may not be possible (ie, would lead to inconsistency of any theory that contains it) or else very particular speculative higher-dimensional physics.
I'm not sure quite what you mean by a bottleneck region. If you can elaborate, I may be able to answer.