[flumpcakes]

An author has lived experiences, including other books they have read, to draw upon to tell a narrative they want to tell (either purely for expression, purely for profit, or more often than not somewhere between).

A machine that chews up the worlds literature and spins out a best guess at what the next word should be does not have intent, and the vast majority of the time is used by unscrupulous people purely for profit and/or deception.

An LLM and a living human being are not the same thing, I am tired of apologists comparing them as if they are.

It's not surprising that a computer (doing trillions of calculations on a billion parameter model that was trained on the world's literature) can string a coherent sentence together...

[rpdillon]

This is predicated on the belief that the AI is running autonomously without ongoing input over hundreds of interactions to produce the output. In all of the professional contexts where I've leveraged AI, every output is the result of hundreds of back and forth interactions, review, modification, and iteration.

Your narrative about next token prediction makes no allowance for this.