That’s really a flawed analogy, though. In this analogy, the man isn’t passing the Chinese Turing Test – the program is passing the Turing Test. The man is just the hardware that the program is running on.

It’s not a question of whether the man (the hardware) has any understanding. What you want to know is if the program has any understanding. Honestly, it seems like deliberately muddying the issue by proposing this hypothetical ‘intelligent hardware’ and using it to distract from the actual question about the software’s intelligence.

If I’m running a supposedly ‘intelligent’ AI on my 3090, you can’t refute that by pointing out that the 3090 is not intelligent on its own, without software.

You gotta send separate emails with only one point/question in each email. Leave them nothing else to respond to.

Gotta burn up the token quota somehow, or the boss will be mad again.

Depending on the humidity in the pit, it could be very dehydrating and cause them to basically mummify.

It will either instantly trip the breaker or do absolutely nothing, depending on how it’s wired internally.

The real danger comes from when only one end is plugged in, though. With only one end plugged in, the other end becomes a 120V cattle prod.