I've been thinking about this and haven't found any fundamental difference. Sure, LLMs don't have our fine temporal resolution, but GPT-4 (at least) knows the

I've been thinking about this and haven't found any fundamental difference. Sure, LLMs don't have our fine temporal resolution, but GPT-4 (at least) knows the date, and can get the current time using Python when asked. And can tell the order of text events within a session. Our resolution has a limit too, somewhere under 1/25 of a second while awake, with much larger gaps when we sleep. So it's a matter of degree, or more to the point it's how we might gerrymander definitions to suit us.