Talking pretty, too -- which takes some intelligence, but not too much. The era of the awkward nerd is over. The world is returning to its natural order, with the slimy salesman in the lead.

Price of human intelligence is going up actually, because there's ever-decreasing amount of it, and we've invented a cheap substitute that makes people even stupider.

Well, someone. But not necessarily anyone else.

Why would it freeze if they can create more training material. Its already been shown synthetic training data is just as good if not better.

Surprisingly, this isn't necessarily true.

To take an intuition: there are lots of problems that are easier to pose and to check than they are to solve, in mathematics.

No, they can just solve tasks verifiably. Reinforcement learning from verifiable rewards. Pure next-token supervision is only in the pre-training phase for modern LLM agents. They can also define new RL environments and pose new challenges to themselves and train themselves to solve them faster. Yes, at some point some human steering and judgment comes in as to what sorts of tasks to train towards. But there is no obvious taper-off at human level.