> So like, there are kids in Africa getting new for of substandard bad education instead of old for of cheap education. Congrats.

Yikes. I don't think you understand Africa at all. The question isn't "a crappy calculus tutor in a rural village or no calculus tutor". The question is "a good but imperfect tutor or nothing". It's VERY different.

> Problem is that you are imagining the collapse of barriers that is simply not happening in economic data. Those barriers were weak 25 or so years ago and were going up last years.

Economic data is a trailing indicator. It would be shocking to see modern frontier models reflected in the 2020-2025 data, right?

But you're right it's a qualitative opinion. But FWIW execs at Fortune 500's are terrified, as the codebase itself is no longer a barrier. This is not a secret; it's being said in all-hands meetings.