It’s a philosophical question, and I personally have very little interest in philosophing. LLMs are technically limited to what is in their training dataset