The immediate hot take is even if correct, the paper is largely opinion piece wrapped in an academic paper with Stanford logo.

Most of the sections had no citations and anecdotes.

Like many who have predicted doom or ChatGPT can never do XYZ, anecdotes do not build a substantive argument.

...What? Like half of the vertical length of the paper is footnotes with citations. Also, opinions are kind of an important part of how we get ideas.