Good idea but...
...how do you get that level of positional detail? GPS is nowhere near. And not only does it need to gather your exact location to the nearest centimetre, it needs to know what direction you are pointing in, what angle you are looking at, and what the field of view angle is, to be able to tag things correctly. Oh, and it all needs to work indoors as well as out.
And most tags will become immediately redundant - there's no point in tagging a specific product on a shop's shelf, because within a week there might be different products.
By the look of the video, it is using artificial intelligence to recognise items by the photo, as well as positional data; being able to run this kind of software on a handheld device is a long, long way away.