Grokipedia is dead. I'd like to rescue it, starting with scientific data

An investor pitch on X. Diagnoses three problems, LLMs ingest and repeat wrong information with certainty, peer-reviewed research has bad incentives and poor results, and much public harm comes from bad science, and claims the only fix is an open two-sided "validation market" where participants make money when right and lose it when wrong; distinguishes this from a prediction market and calls it better than Community Notes. No mechanism for resolving what counts as right is given.