Comment by the_af

I don't think I did.

If you follow the spirit of the essay, Ted Chiang suspects Anthropic is being cute with the idea. There's no good reason to suspect consciousness in LLMs, so the null hypothesis must be taken as default. And Anthropic sort of knows this, but for marketing purposes they are playing loose with definitions, hedging their bets and drawing up "constitutions" with terms for the "well being" and "happiness" of Claude, while at the same time -- and this is an important part of the essay -- being unethical (think slavery) if we assume they truly believe Claude could be conscious.

Of course Anthropic is being cute about this, they have a vested interest in hype and overpromising; even drumming up the "AI danger" is a way of hyping up the tech.

Ted Chiang is taking the default and honest position: "no, LLMs aren't conscious. If you truly believe they are, show us the really scientific and rigorous proof."

In practice this is a much weaker stance than saying "maybe they are conscious". It is the only honest scientific stance, really.