Watched it. Charles Stross talks about the AIs already among us, who were implemented in the 18th century and gained legal personhood in the 19th century, and notes how blind many people warning against paper clip maximizers are to the fact that we already have revenue/eyeball/outrage maximizers.

I like that he brings up the Chinese Room in the Q&A. In my view, the Chinese Room intends to disprove AI, but the room *is*, in fact, intelligent. And the same goes for a corporation.

A corporation is an AI implemented on a human/paper/computer substrate, and its computer component, and therefore effectiveness, is increasing.

What Stross himself fails to see is that his countermeasure, the state, is an even older AI[0], which does its own maximizing independently of the humans that power its machinery and the overt goals set by its voters and leaders.

[0] or -- if you count only the impersonal liberal democracy, as most older states were more, but not entirely, like a prosthetic of the individual ruler -- about as old as the corporation