Key information
The indie title Teach My Little Sister How To Drive has encountered severe financial difficulties following the viral success of its demo in Asia and Japan. Developed by the four-person studio Easy Fox, the game tasks players with instructing a virtual sister via AI to navigate driving challenges. Despite its popularity, the project is currently incurring significant debt.
The game utilizes a complex AI architecture, relying on Google Gemini 2.5-flash as its primary engine, with OpenAI's ChatGPT Realtime Mini 2.1 and ChatGPT Realtime Mini serving as backups. To maintain thematic consistency, developers implemented primary prompts to ensure the AI remains focused on driving, though they acknowledge the system is not infallible.
The studio reports that the high volume of players has caused operational costs to surge, with daily expenses for AI services exceeding 1,000 dollars. Because the game relies on online token-based requests, the viral interest has strained the team's budget, forcing them to secure a bank loan to keep the demo accessible to the public.
Easy Fox is now considering transitioning to local AI models to mitigate costs. While the demo remains available ahead of the scheduled January 2027 release, the team has contemplated temporarily closing the trial to manage their financial situation while they seek a sustainable solution for their infrastructure.
