On Monday, 13 July 2026 at 21:37:26 UTC, H. S. Teoh wrote: > I thought the whole point of an LLM is to replicate patterns > from data in its training set? :-D People see what they want to see :-)