In Lesson 1 you have "Current models are thought to be trained on terabytes of text and code". That's surely too small. You can get a 1TB on an SD card. Lesson 10 mentions "petabytes of overwrought fan fiction". I suspect you may need to go to exabytes.
Comments