Comment by rany_
3 days ago
This is really interesting. How can you detect when it's the same person passing a captcha? I don't think IP addresses are of any use here as Anti-Captcha proxies everything to their customer's IP address.
3 days ago
This is really interesting. How can you detect when it's the same person passing a captcha? I don't think IP addresses are of any use here as Anti-Captcha proxies everything to their customer's IP address.
I don't know exactly what they do now, bloom filters was a thing then, also lots of heuristic approaches based on the bots we detected. the OP agent example actually would fail the very first test I deployed which looked for basic characteristics of the mouse movement
Here's a fun experiment for someone: 1) Give N people K fake credit cards to enter into a form, and have them solve a captcha 2) Take recorded keyboard and mouse data similar to the captcha 3) Train a neural network model to identify
I've been out of this for 6 years but I bet transformers rock this problem now.
TIL that keyboard and mouse data is used to identify "bad" humans not just bots. It's a bit freaky but not at all surprising.