Google employee here. One of our updates to Captcha started using context-sensitivity using a neural net to process keywords in webpages to determine where to send 'select X image' problems. For example, a forum or blog with an audience who are familiar with motorcycles will get more Captchas with motorcycles. The more weight the algorithm assigns to the usergroups experience and knowledge, the more likely it is to use puzzles with image recognition involving those things.
For a while now you've been annoying the fuck out of everyone in the Captcha dev team because you keep skewing the metrics. The Captcha system is convinced, via neural net deduction, that the entire *.Sup Forums.org userbase are experts on helicopters and streets.
Michael Cook
...
Leo Cook
It freaked us out in the early pizzagate threads, the capchas started being pizza images, we figured out thats what it was.
Jonathan Martin
Haha ha hahah hahaha
Joseph Nelson
seems legit
Luis Wood
Have never seen a helicopter captcha in my life.
>They dont exist
Adrian Foster
Hello Google employee. I have two things to say to you.
1. Fuck you. 2. Are you hiring?
Blake Young
Okay
Bentley Howard
It gets better. The more captchas people successfully solve on a subject, the more likely our search engine is to rank that domain or webpage on the domain when people are searching for that subject, OR, and this is the fun part: KEYWORDS ASSOCIATED WITH HIGH RANKING DOMAINS OF EXPERTISE ON THAT SUBJECT
The neural net started filling search results for Pizza with Pizzagate links. I'm pretty sure someone got fired because of that.
