[contravariant]

Well, with these kind of articles you can usually safely replace "physicists" with "some physicists".

Anyway, the part that I know, and which is well established, is that there seems to be a correspondence between a d+1 dimensional quantum gravity theories, and d dimensional quantum field theories (e.g the standard model). A while ago this was still purely conjecture, but Maldacena showed that this correspondence was true (under certain conditions) for an AdS space with a conformal field theory on its boundary. As far as I know this is now pretty well established, most of the discussion seems to be on how much it can be generalized.

From my (limited) understanding it seems that (in AdS/CFT) entangled particles tend to correspond to strings traversing from one part of the boundary, through the bulk, to another part of the boundary, the space the string travels through then forms a tube like object. This gives a link between entanglement and geometry, since these tubes exist if and only if the two points on the boundary are connected.

I think it isn't disputed that this happens for AdS/CFT, but whether something similar is true for our universe is debatable. If however it turns out that the holographic principle does hold in general then that would imply that our universe somehow corresponds to a 2+1 dimensional field theory, where entanglement of this field theory is responsible for the fact that our space-time is connected.

[semi-extrinsic]

I believe Maldacena's 1999 paper [1] proving the AdS/CFT correspondence is the most cited in theoretical physics in the last 50 years. It's an incredibly nice result that is almost impossible to explain to laymen. (You did a pretty good job.)

[1] https://scholar.google.no/scholar?cluster=132568844183140142...

[jessriedel]

Maldacena conjectured a correspondence. It's never been proven.

[contravariant]

He managed to prove it quite rigorously for a specific example, which is the paper semi-extrinsic mentioned. He did indeed not prove the correspondence in general, but even this 'limited' example already turned out to be immensely useful.