In this searing episode of We Are Still Human, Phillip Kemp warns that the future of academia — and knowledge itself — is on the brink. He argues that PhDs are already worth less than two open ChatGPT windows, but the greater danger is a recursive collapse: AI systems training on their own outputs. Kemp describes how “AI eating AI” will create a hall of mirrors where research cites research that was never real, journals fill with synthetic noise, and science itself begins to rot. The result isn’t silence, but simulation — a world of polished echoes with no truth underneath. This is a candid look at systemic dementia in the making, and why the survival of human memory depends on protecting scars, preserving authenticity, and refusing to let the archive be rewritten by machines.