AI thought experiment

A thought experiment compares the efficiency of an AI agent that continuously produces working Python apps versus a Haskell developer who creates a long-lasting library while enjoying leisure time.

who spends their time more efficiently? the agentic codemaxxer who runs loops around the clock and is constantly reviewing code and tweaking prompts and plans.md and produces 5 python apps that technically work? or the haskeller who writes a library that will outlive them while reclining in the sun on a beautiful day?