[trc001]

Uh, the bellman equation was first used for control theory and is the foundation of modern reinforcement learning... so wouldn't that imply LLMs "come from" control theory?

[fc417fc802]

Is the training algorithm the AI or is the model that you get at the end the AI?