The editors of Nature should know better than to put a headline like "faster-than-light information transfer" on there, since that is exactly what isn't going on.
Um, this part seems to imply so... "the researchers found that when each photon reached its destination [some 18 kilometres apart], it could instantly sense its twin’s behaviour without any direct communication."
The key is, this "sensing its twin's behavior" only works if you interfere with the twin in a randomized way. You can't set one particle to a 0 bit and have the other one sense it. So you can't use this for sending information.
Think of ‘entangled’ photons as sharing a secret. If either one gives up that secret than you also know what the other one is doing. The only spooky part they don't behave like they have a specific value until you measure it and you can randomly change it's value.
PS: You can't send information from a to b using two ‘entangled’ photons sent from c and read at a and b.
Um, this part seems to imply so... "the researchers found that when each photon reached its destination [some 18 kilometres apart], it could instantly sense its twin’s behaviour without any direct communication."