Comment by kypro

> Instead, I asked Fable to explain [TC26a] to me. Fable seems to understand the argument and thinks that it is correct. I expect that thousands of frontier LLM instances around the world have come to the same conclusion over the last two weeks. If not, surely we would have heard by now.

This is the future.

Advanced AIs are basically genie's bound only by the laws of nature (which are necessarily the known laws). You ask the genie for something and if it's possible, it will grant your wish. What it does will seem like magic to you or I, and while in theory there is an explanation for the magic it's preforming, in reality an explanation could be as complex as understanding how the AI itself works at the level of its weights. It's not something a human mind can do.

The genie might give you a cure for cancer, but you won't know how it works. It's possible it will cure all cancer but with some hidden cost. A cost we cannot know and one we must trust the genie is both aware of and would inform us of. Maybe it will make us more trusting of the genie, or maybe it will just make us all drop dead in 20 years time.

The world we are now entering will look similar to the world native populations entered when they encountered Europeans. Suddenly people were turning off the sun and yielding magic kill sticks. It was basically impossible to reason about what might or might not be possible.

This future, is for now, still optional. We can still decide if we want it, but when the genie in its final arrives, perhaps one of the only things it won't be able to grant us is the ability to undo time and take back our decision if we later decide it was a mistake.

The most important thing about this paper is that it's describing human loss of control. A world where we are no longer the masters of our own destiny.