Just in case you weren't sure it was nonsense by 1:25, they introduce "secret mode".
Phase cancellation on a window with sounds coming from outside... that's fine. A good implementation would be really difficult because of echoes and reverb and stuff, but it's technically doable.
"Secret mode" however, that's just not going to work. Before the device can emit the phase cancelling sound, your speech has to travel from your mouth to the device. By that time it has already travelled towards other people/surfaces to be heard/reflected.
Anyone listening in would probably hear multiples of your conversation arriving at their ears at different times.
At best eavesdroppers would hear a hubbub of unintelligible conversation, certainly not the silence the video portrays.
If they actually had a workable prototype running on a laptop, even with an elaborate microphone and speaker setup, they would have demoed it.