The way I understand it (And also I'm a laymen, just have an active interest in the topic), the problem with entanglement is that it
is instant. How do you work out a timing system accurate enough to measure a clock cycle when the two computers are on opposite sides of the room, let alone other sides of the world? Remember that a photon is constantly bouncing around at the speed of light. If I'm reading this correctly their accuracy improvement from 25% to (85 to 90)% has to do with correctly differentiating and storing this state change.
https://www.sciencealert.com/scientists-just-unveiled-the-first-ever-photo-of-quantum-entanglement <- The photo here shows 4 states of a photon, in all it shows an experiment to record/manipulate the alignment of an entangled photon (the other half of the pair was in a different "box" as it were). If you look at the photo of the entangled particle, there are outliers, pixels around and near the key structures. So instead of looking at a single snapshot to determine the alignment, they had to take thousands of images to build the structure that shows a general idea of how the photons were aligned. The improvement from 20% to 90% means that the structure can be determined with fewer "pictures", which means that the readings are both more accurate, and faster to compile.
This effects the number of times you have to read and resend the data in order to confirm correct delivery of an information packet, kind of like what we do with tcp vs udp...
I'm pretty sure this same problem was encountered when fiber-optics were in their infancy as well. The computers/ram weren't fast enough to manage the data so new hardware had to be invented. But now we are dealing with a system that's faster than light and requires no physical wire infrastructure to maintain. (Though the energy consumption of these experiments today is tremendous, this should be less each year as we shrink the technology down and increase its efficiency)
In the next decade this might affect the world's information on a government level rather than a personal level. At least until quantum computers are being sold to the general public (maybe half a century from now?) It's in its infancy, but it's like watching the first people experimenting and working out the shape and usage of a wheel, it might end up being nothing or it might change society as we know it...
I have edited this so many times in the last hour that I've lost track. I apologize, but this had me brimming with excitement.