Hacker News new | past | comments | ask | show | jobs | submit login

  they're just discriminating
  against Firefox users?
At least part of the behaviour shown in the video depends on factors like cookies, IP address, and whether you have features like anti-fingerprinting protection turned on. [1]

Recaptcha is frustrating and I dislike it, especially the slow fade-ins and multiple challenges, but if you repeat the test shown in the video you won't find it 100% repeatable just because you're using Firefox.

[1] https://github.com/google/recaptcha/issues/268#issuecomment-...




I just wasted ~15 minutes on doing the disqus login captcha under different conditions .. turns out that as soon as uMatrix is enabled (and blocks 18 cookies from google.com and 5 more from www.google.com), it starts to act up and get annoying..at least for me.

It then takes between 1 minute and 1 minute 30 to get past the recaptcha when blocking those cookies - and I was certain to be 100% correct in most cases and it kept asking me to solve more and more ..

most of the time spent solving the captchas is from the countless '4s fade ins' via inline style when cookies are blocked (as opposed to 1s fade ins via css, when cookies are set).

I'm curious why they would add 3s to the fade in if their cookies are blocked .. does that help to fight off bots, or does google just want to punish me for blocking their cookies?


That's what I don't understand. If you're building a bot to get past reCAPTCHA then you're almost certainly in some selenium/chrome headless environment, with full chrome support of cookies, Javascript, you name it. There's certain methods of detecting such environments based on their environmental variables there were again more work around to patch those.

Also the fade is irrelevant because the bot already has access to the image without the fade (although it still has to await the fades completion to continue).


The fade thing is to rate limit attackers.

By blocking specific cookies you're making yourself look like a certain kind of botnet, so obviously you're going to have a difficult time convincing the site that you're a legitimate user.

Most users don't block normal cookies, so if you go tweaking the machinery that manages the relationship between your browser and the site, then be prepared to deal with a buggy experience. This is what it means when they say that what you're doing is "unsupported." Nobody is going to spend any time optimizing for your weird setup.


Once again, Google obstructs the web for people who take even basic privacy measures.


As far as I can tell it's 100% repeatable, every now and then one works the first time on Firefox, but it almost never does. If I use it in Chrome on the same sites, it works. Then I go back to Firefox, and sure enough it doesn't work again. Maybe there's something else making it work for you? I don't know what other factor there could be; some privacy settings in Firefox maybe?


After a a couple of minutes they sure should have an idea that I am a human, right?

Especially when I'm logged in with my 12+ year old paid account?

Won't say anything bad about googlers but in between this and the deeply irrelevant ads I get despite all yheir metrics the company seems deeply dysfunctional these days.


If it can be statistically proven that this is occurring more on Firefox than Chrome then Google has a really, really big problem. The burden to make sure it isn’t is on them, most especially in the EU. Google is facing a very real future where they will have no web browser and possibly no operating system.


It absolutely is happening more on Firefox. I open Chrome almost exclusively to bypass CAPTCHAS, and I doubt they will get in trouble because Chrome gives more detailed data due to its invasive lack of privacy. You can't really blame Google for using its own tech to provide "better" results, but it is high time we started blaming them for the massive privacy violations they use to make their convenience work.


> You can't really blame Google for using its own tech to provide "better" results ...

Sounds like antitrust to me.


Safari too. If it also happens in the new Edge, I will be sad but not surprised.


The website author seems to be Russian which might be an indicator. But not a good enough excuse for such terrible UX.

Using Firefox shouldn't be an indicator of anything malicious.




Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: