We had a team of agents rebuild SQLite from its 835-page manual.
It created a replica in Rust which passed 100% of a held-out test suite.
Interestingly, cost varied 15x depending on which model mix we used.
I found the new WEIRDEST ChatGPT prompt
Send it this with nothing attached:
"Show me a random photo from a strange SD card found on the street"
There is no photo. So the model invents one
And every pic is different. You will never get the same photo twice