Veggies for cats and dogs
« previous post | next post »
This video was passed on by Tim Leonard, who remarks, "real-time video translation at its best":
The Japanese in the background reads:
inu / neko no taberu nama yasai 犬・猫の食べる生野菜 ("raw vegetables / greens for dogs and cats to eat")
*nama 生 ("raw) > shinsen 新鮮 ("fresh")
You can see for yourself what the English translation is.
[Thanks to Nathan Hopson]
Jenny Chu said,
May 1, 2017 @ 10:28 pm
The banner at the top – "prevent gastrointestinal disorders by hell good" – is almost as extraordinary as the main attraction.
Victor Mair said,
May 1, 2017 @ 10:28 pm
From Nathan Hopson:
The description above is out of focus and cut off, but it looks like the bottom line is:
胃にたまった毛玉を、はき出させる効果があります。
(I ni tamatta kedama o hakidasaseru kōka ga arimasu)
"Causes [pets] to vomit up hairballs"
Or more literally:
"Has the effect of causing [dogs and cats] to vomit up hairballs built up in their stomachs."
I can't make out any more than that, and neither can this Reddit thread…
Jenny Chu said,
May 2, 2017 @ 2:01 am
Should we be concerned that Reddit forces me to swear that I am 18 and ready for NSFW content before viewing it … while LL makes no such warnings? :) :)
David Morris said,
May 2, 2017 @ 3:36 am
I've been shopping at the wrong supermarket all this time!
Prosper Noah said,
May 2, 2017 @ 7:16 pm
I can't really make head or tail of what this video is about!
Gene said,
May 2, 2017 @ 10:14 pm
That smart Phone is a cunning linguist
Rodger C said,
May 3, 2017 @ 7:11 am
Gene wins the Internet!
Mark Metcalf said,
May 3, 2017 @ 2:06 pm
It's a mobile, camera-enabled version of Google Translate. Just installed an iOS version on an iPad and tried it with some Chinese and Japanese texts. It performs just like I expected a mobile camera-enabled version of Google Translate to perform (i.e. see the video above). At least the price was right.
Here are more details on the app: http://appleinsider.com/articles/17/01/26/google-adds-instant-japanese-translation-to-translates-iphone-camera-based-word-lens