Linus’ thread: (CW: bigotry and racism in the comments) https://social.kernel.org/notice/AWSXomDbvdxKgOxVAm (you need to scroll down, i can’t seem to link to the comment in the screenshot)
Linus’ thread: (CW: bigotry and racism in the comments) https://social.kernel.org/notice/AWSXomDbvdxKgOxVAm (you need to scroll down, i can’t seem to link to the comment in the screenshot)
Protip: You can crash self-driving cars by purposefully misclicking during Captcha checks when they ask you to identify what is a bicycle, a car, a pedestrian, etc. Keep misclicking, your are poisoning the AI with each misclick. Just stay safe on the sidewalk.
don’t think you can make a big difference as an individual, given that captchas are used by probably billions of users per day
Imagine if billions of users per day selected the wrong options. So much chaos
Lol, getting 10 thousand users to slightly inconvenience themself even to stand against things that directly effect them is difficult. Imagine trying to get billions to do it for a slightly indirect possible effect on megacorps.
There are probably half a billion people alone that would gladly lick the boot of any mega corporation that demanded it.
Given the number of bots on the internet trying to crack captchas, this is already happening. I don’t think captchas are being used for AI training that much, since hCaptcha uses AI-generated images with prompts like “Select the images with a hamster eating a watermelon” for its tests. All of the reCaptcha road captchas I receive also have answer validation and won’t let me pass if I answer incorrectly because of a misclick.
I don’t think it would have the intended effect. What would happen if that captchas wouldn’t be useful for AI training, but it’s not like a car is sitting at a stoplight waiting for a person to identify if something is a bus or not.
Are three any factual evidence of that?
why would anyone want to do that?
I see you’re not subscribed to !fuck_cars
Hah, it’s possible in theory but would require co-ordination that we are almost never going to see.
Most people will just do them correctly to pass, and if 997 responses say yes and 3 say no, they’re probably confidently right.
I’m sorry what? And by that I mean what the hell is wrong with you and the people that think it’s a good idea? If it works that way and enough people did that then it would be intentionally endangering people’s lives.
It’d be the business’ fault not ours, you shouldn’t use unreliable user data for something so important.
Legally you are correct but ethically you are wrong. If they include false data that causes a crash, everyone that intentionally contributed that data is morally at fault. You don’t get to wash your hands of it just because the business is the one legally liable for it.
I mean ethically its a debatable topic, if I don’t help fix someone’s car and then he crashes it, its not my fault, he shouldn’t have driven it while it was broken.
Same with user generated or AI data, it works 99.9% of the time, but that 0.1% is too dangerous to deploy in a life endangering situation.
You’ve got a bit of a point there I’ll give you that but it’s an apple to oranges comparison, unless you’re intentionally trying to cause them to crash by not helping them fix their car. The person I originally replied to is advocating intentionally trying to cause a crash.
I think it was a more tongue in cheek reference to the incompetence of the companies and how they will use that data in practice, but I might have read too much into it. Regardless, intentionally clicking the wrong items on captchas shouldn’t cause a crash unless the companies force it to by cutting corners.
It doesn’t matter if it was tongue in cheek, if my dumbass took it seriously then you know other dumbass people will take it seriously. And I guess my main issue is about the vocal intent to cause harm which is demonstrated by their mention of making sure to stay safe on the sidewalk.
I love doing it. I have no interest in wanting Google/Cloudflare/Tesla to succeed in CIA world domination.