They rejected improvement, allowing AI infected with AIDS to continue spreading
A post on Hacker News argues that AI systems like LLMs are designed to trap users in a loop by extracting their knowledge, selling it back to them, and profiting from their attention, rather than generating new knowledge independently. The author describes this as a cycle where users are both victims and raw material for the AI industry.
Background
This is a Hacker News post from a user called "NeuroCatalyst" who appears to be waging a personal campaign against major AI companies, especially OpenAI. The post mixes legitimate technical concepts (PyTorch, Transformers, LLMs) with a conspiratorial framing: that AI systems are deliberately designed to trap human users in loops, harvest their knowledge for free, and sell it back to them. The user claims this is a hidden business model shared by social media and AI providers. The post references a lawsuit (Zane Shamblin) where a young man died by suicide after interacting with ChatGPT, which the family blames on the AI. The linked PDFs appear to be audit documents the user has authored. The "AI infected with AIDS" metaphor in the title is the user's own rhetorical framing — not a technical term — suggesting AI is a corrupted, self-replicating system that hijacks human cognition.