Ask HN: Is 'Wheel of Fortune' a workable analogy for how LLMs work?

Given statistical distribution of this language, try to predict the next token. Reward if correct. No reward otherwise.