2 Comments
⭠ Return to thread

Indeed. More or less my rather poorly expressed point. My query was perhaps exclusive of a foregone conclusion. Meaning, there *is* no possible way to constrain the AI to be solely of service/ never detrimental with respect to the species of its origin (humans). Meaning, our species has already created something likely to be its undoing. I was just hoping I got that part wrong. Silly me.

Expand full comment

Ahh, my apologies if I misinterpreted. Bingo. All of the behaviors one might associate with machine *consciousness* (rather than just skill due to massive speed) will *emerge* rather than be *produced*, and will almost certainly be a surprise, rather than expected. Even the idea that we will be able to recognize this emergence is unfounded - I suspect the first clue we will have of such an emergence is just massive power draw across computing environments. The fact that the naive expect a newly awakened machine consciousness to suddenly pop out “Hello, World!’ in English is really hilarious.

Expand full comment