undefined | Better HN

0 pointsadventured5y ago0 comments

Google intentionally tamped down their reverse image search several years ago. They fundamentally changed how the product works, the results that it delivers, and did it on purpose. It's definitely inferior if you're actually looking to find copies of an image (Google is fully aware, they don't want you to be able to effectively search their image system that way).

0 comments

5 comments · 2 top-level

pizza5y ago· 2 in thread

Interestingly enough, Yandex did so also, a while ago, as it was being used as a scarily good open-source undercover agent identification tool https://www.bellingcat.com/resources/how-tos/2019/12/26/guid...

although I can't seem to find the followup post that actually discusses how the Yandex search quality decreased following the publication of the original article

llampx5y ago

Fascinating read, thank you. I can't help but wonder what is available to big companies and governments these days in terms of technology and algorithms that we laypeople don't have access to or even dream about.

exikyut5y ago

Ah, wow, finally found the exact article I was looking for. This thing: https://news.ycombinator.com/item?id=22492671

And then there was the genomics or DNA database or something that a couple of people were running out of their garage (well, a bit more than that) but then had to shut down after law enforcement discovered it, found it incredibly useful, and everything rapidly went sideways as "things aren't supposed to work that way"... but I'm having the hardest time finding the link for this one sadly ._.

EDIT (just within the 1hr window :): I think the second one was GEDmatch. A quick search of https://hn.algolia.com/?query=gedmatch didn't relocate what I remember reading, but I think this was it.

beefield5y ago· 1 in thread

> they don't want you to be able to effectively search their image system that way)

Why?

dvfjsdhgfv5y ago

This is an excellent question.

j / k navigate · click thread line to collapse