Comment by narag

Seeing this here again, I was wondering if it's related to current events. I hope it isn't too soon, sorry to bother you in that case.

Can't you see? Let me explain.

The point of the post is that when you try to translate ideas to the physical world, there are plenty of quirks and you quickly find out what's the difference between theory and practice. That's why many people think that the chatting performance of an AI isn't a good measure of its ability to conquest the world.

And that's why some are so fed up with the doom prophecies, they seem disconnected from reality. But I believe the general opinion has swung too much in the other direction and maybe, just maybe, it's not wise to completely stop thinking about safety.