GitHub's explosive growth to 14 billion commits annually reveals AI's true economic value lies in closed-loop, infinitely-demanded tasks like software engineering.

GitHub experienced massive scaling in 2025: 1 billion total commits, 275 million commits per week, and GitHub Actions grew from 500M to 2.1B minutes weekly

Kyle Daigle, GitHub's COO, presented a 2x2 matrix categorizing AI work by demand ceiling (infinite vs. finite) and loop closure (closed vs. open)

Closed-loop infinite demand tasks like software engineering are 'economic engines' where AI writes code, automated tests verify correctness, and continuous output creates compounding value

Closed-loop finite demand tasks like AI bookkeeping are 'efficiency plays' with limited economic upside despite automation benefits