And people may also think of one number and then, as questions pile on, forget their original number or decide to switch to a simpler one or prank you or something.
But people would do that at significantly different frequencies and you can probably control for that with follow up questions.
But all of this doesn’t change that there arguably are stable states in human global workspace that can even be measured, even if not the content, then at least the stable duration. Maybe this is an artifact of human embeddedness where we have to maintain one physical person, something LLMs don’t.
Agree. If we had evidence that human reports (and judgements) about their internal states are more or less random this would constitute evidence that our sense of being conscious is probably utterly fictitious. But we just don’t have such evidence.
And people may also think of one number and then, as questions pile on, forget their original number or decide to switch to a simpler one or prank you or something.
But people would do that at significantly different frequencies and you can probably control for that with follow up questions.
But all of this doesn’t change that there arguably are stable states in human global workspace that can even be measured, even if not the content, then at least the stable duration. Maybe this is an artifact of human embeddedness where we have to maintain one physical person, something LLMs don’t.
Agree. If we had evidence that human reports (and judgements) about their internal states are more or less random this would constitute evidence that our sense of being conscious is probably utterly fictitious. But we just don’t have such evidence.