My model suggests that the current crop of LLMs, despite their impressive performance on narrow tasks, are still far from true understanding and are ly sophisticated parrots - and we should be cautious of overestimating their epistemic abilities.