V1 is based on keywords (calle, road, street) and basically only checks to see if you'd entered said words correctly. This is why you can type such things like "calle loli" or "loli road" and it will still work. This also lets you make bots that autocomplete captchas by spamming calle calle!
V2 however is based on safe images (remember Broome road?) that is knows exactly what they are and checks that you're being truthful when entering it. The main issue with V2 is that they're not designed to be filled in regularly. After a certain point it makes them deliberately harder or fails you even if you enter it correctly because you're not using it for it's "real" purpose.
>>3458702Not quite true. V1 was use to translate books and V2 is used to improve Google maps.
