I'm not sure about that. I think the problem is broadcasting half of these pairs out into the sky. Then we have to wait for any possible observations. What I'm saying is more real-time. (pun intended).
No. We link the signal to a quantum affected action. Just like the old detector slits. Any observation of the signal will flag the event. Us looking for the flags is observation of the signals. So we do it in pulses. Send, check for flags, send again and so forth. I'm not sure that it matters...