i think their point was that people do this stuff to hide it from algorithms so that it doesnt get removed by the platform, when algorithms can easily read the covered text anyway
And if anyone thinks social media networks aren’t doing this and more are for a rude awakening
If they need to do a chatgpt scan of every single image that is uploaded, that’s gonna cost a lot of time and money. They just to OCR and the OCR can’t read that
It does literally nothing. The free version of ChatGPT decoded this advanced encryption in seconds
And if anyone thinks social media networks aren’t doing this and more are for a rude awakening
I deciphered it in milliseconds without having to paste it into a digital slave
i think their point was that people do this stuff to hide it from algorithms so that it doesnt get removed by the platform, when algorithms can easily read the covered text anyway
The point is that the companies use “digital slaves” to decode this preschool level “censoring” in milliseconds, so doing it does nothing of value
Their systems will now know that the user posted about Epstein, despite the red line over the name.
If they need to do a chatgpt scan of every single image that is uploaded, that’s gonna cost a lot of time and money. They just to OCR and the OCR can’t read that
The don’t “ChatGPT”, you can use local models and they cost next to nothing on Meta’s or Google’s scale. Both run their own servers for it.
of course they don’t use chatgpt and whatever they use isn’t comparable to chatgpt cuz that would be unsustainable
“Isn’t comparable”? For generic tasks that’s true.
Figuring out shittily censored words from pictures and subtitles? The custom models are even better