I suspect that fact that the vast majority of LLM users don’t seem to have received this (really very simple!) message is because the AI companies have a vested interest in us not understanding it. “ChatGPT is smart, it just makes mistakes sometimes” is much more marketable than the truth.