In July 2026, Ethan Mollick showcased how AI coding assistants can radically reinterpret existing game code when he prompted an advanced model called Fable to transform his earlier GPT-5 procedural brutalist city builder into an entirely new experience titled DEEP TIME. The original brutalist city builder demonstrated GPT-5 level code generation for urban layouts. Within less than a year, the same prompt strategy applied to newer models produced DEEP TIME, a simulation where players build a city, observe its abandonment over simulated centuries, and then excavate it as future archaeologists. Modern AI models interpret high-level creative instructions such as "add abandonment and archaeological discovery phases" and output functional code that integrates new mechanics like decay timers and excavation tools. Game studios can leverage these tools to prototype multiple variants of a single base project in hours instead of weeks, opening monetization paths through niche simulation titles and educational partnerships with museums. Publishers may release base tools powered by AI and sell expansion packs focused on specific historical periods. Continued improvements in AI code synthesis will likely accelerate the creation of complex narrative-driven simulations, shifting industry focus toward curation and player experience design rather than raw implementation.