👏 Poison 👏 your 👏 data ☠️
-
@alice I've toyed with the idea of setting up a headless Chrome instance to just ask "but why?" to ChatGPT all day to drive up their inference costs. 👀
@theorangetheme ha! thats so good.
-
@alice I've toyed with the idea of setting up a headless Chrome instance to just ask "but why?" to ChatGPT all day to drive up their inference costs. 👀
@theorangetheme I'd love to make a bot that just hits all the big LLMs and corporate "help" bots, and simply shuffles their answers to each other randomly, then randomly up or downvotes the replies.
-
👏 Poison 👏 your 👏 data ☠️
@alice bonus points if you have a #MeatProxy !
-
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
@alice when i have to use a web app to order food, e.g. CoolBurgz (fictional) i will always put my email as e.g.
coolburgz@coolburgz.coolburgz
usually counts as valid.
-
@alice when i have to use a web app to order food, e.g. CoolBurgz (fictional) i will always put my email as e.g.
coolburgz@coolburgz.coolburgz
usually counts as valid.
-
@theorangetheme I'd love to make a bot that just hits all the big LLMs and corporate "help" bots, and simply shuffles their answers to each other randomly, then randomly up or downvotes the replies.
@alice @theorangetheme I once built a fuzz testing tool that "randomly" shuffled input around and tested it against things. "does my input validation survive utterly batshit inputs?"
Feeding the inputs through something like that would make sure they can't cache answers.
-
-
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
a fair bit of the advice in here seems really good, but from what I know, AdNauseam isn't really worth using over just uBO
at least as of when I last looked into it a couple years ago: it uses more resources on your machine, doesn't really make any significant difference for the companies, and the high volume of "clicks" from you just makes you far more trackable since no normal person browsing would do so
also, I think it might be worth editing the last point to say "hopefully none of you are using LLMs, but if you're someone who does..." 🩵
-
@Irenetherogue sure! There are low tech ways to do it—just lie...to every corporation, app, and marketer you can. Make it plausible, but wrong.
Bonus: include something wildly implausible once in a while. It makes folx more likely to overlook the subtle ones.
@alice @Irenetherogue I got off when taken to court for nonpayment of Poll Tax (Thatcher thing, yes, I'm that old) because I poisoned their data by missing out a crucial box on the form.
Don't refuse to comply but *always* sabotage their data. It's simply costs them more.
-
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
@alice "Fold your punch cards"! 😃
-
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
@alice NULL is also a good answer for when you don't want to give out a particular personal detail.
Aside from phone, date of birth, and email, most of the time the front end form fields will accept NULL as an answer.
https://en.wikipedia.org/wiki/Null_(SQL) -
👏 Poison 👏 your 👏 data ☠️
Sorry accidentally poisoned cuzco instead
#With-the-poison #the-poison-for-cuzco #the-poison-made-specifically-for-cuzco -
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
We should tax corporations by the GigaByte of storage the own.
It doesn't matter what they use it for, it should have a tangible yearly cost, to make them think about how much they store.
-
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
@alice Enter your name as [object Object] and let them try to find a bug.
-
👏 Poison 👏 your 👏 data ☠️
@alice a friend of mine changed his middle name to 'undefined', it caused so many problems he had to change it back within a year.
-
@alice I've toyed with the idea of setting up a headless Chrome instance to just ask "but why?" to ChatGPT all day to drive up their inference costs. 👀
@theorangetheme @alice lol somebody has a toddler
-
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
@alice thank you! I've always wondered whether to put random made-up data but hearing the reasoning and logic spelt out like this is convincing me to actually start. Especially commonsense things like "mess with the fields that don't matter in their service to you"
-
@alice a friend of mine changed his middle name to 'undefined', it caused so many problems he had to change it back within a year.
@TheMightyGit @alice
Reminds me of someone who had the last name "Null" and had similar problems. -
The goal is to make corporate data less profitable.
Even stuff as simple as setting your birthdate to 1970-01-01 everywhere, adding [TEST] or [DELETED] as your name or account notes anywhere you don't need them to know your name.
Using plugins like AdNauseam to poison ad trackers (and cost them marketing dollars).
Using VPNs set to different locations.
Signing into data broker sites to "correct" outdated info (they'll often let you do that with little-to-no proof of identity, but will require your passport or state ID in order to delete your info). Bonus points if you correct it to someone else's info on their site that's similar to yours.
Only fill in required fields when you sign up for anything, but only provide correct info if it matters for you to use the service, otherwise provide plausible, but incorrect, data.
If you use LLMs anywhere, use the free tier and always vote thumbs up for bad answers and down for good ones. It wastes their resources and drives up their costs while making their training data worse.
Wrt #PII, It might be a good idea to avoid entering data easily identifiable as trash, and use generators instead. E.g.:
-
@alice I've toyed with the idea of setting up a headless Chrome instance to just ask "but why?" to ChatGPT all day to drive up their inference costs. 👀
'and then?'