I just saw a YouTube video that pointed out how LLMs could “collapse” quite fast if you keep feeding them their own input or gibberish data even in relatively small amounts.
https://youtu.be/3NAYhyuVQk0?si=VptT00iQGOOmN\_bF
There are movements that work to release poisoned data onto the internet so the next generation of LLMs would ingest it during training.
While poisoning the data requires you to be an active participant in the movement, you could still help passively by NOT contributing real human data onto the internet right now. Especially to sites like Reddit (like this post), because they will inevitably be scraped for training. You could still use services and consume existing media but consider not contributing authentic data for LLMs anywhere. Doing this will mean that the ratio between authentic data and AI generated data will continually decrease, and models will be trained on their own outputs from a generation ago.
I’m not gonna naively assume everyone on the internet is gonna mobilize soon because there’s not enough awareness yet. But there’s a very real fear in me and many others that many of us will lose our jobs. If the situation gets desperate, this is legitimately the easiest way YOU can fight back against the Epstein class and their trillion dollars of investment.