High school student creates website to evaluate AI models using Minecraft
A high school student named Adi Singha created the Minecraft Benchmark (or MC-Bench) website, which uniquely uses the sandbox game Minecraft to evaluate various AI models.
A high school student named Adi Singha created the Minecraft Benchmark (or MC-Bench) website, which uniquely uses the sandbox game Minecraft to evaluate various AI models.
MC-Bench offers an intuitive and fun way to evaluate artificial intelligence models. Developers feed various clues to AI models, which then generate corresponding Minecraft structures. Users vote for the best result without knowing which AI model created the build. Only after voting do users see the AI creator. This “blind voting” mechanism is aimed at more objectively reflecting the real capabilities of AI models.
Adi Singh says that Minecraft was chosen not only because of its popularity, but also because the game's visual style makes it easy for even non-gamers to tell which block-based object looks more realistic. He believes that Minecraft makes "progress in AI development more visible" by offering a more compelling visual assessment than purely text-based metrics, TechCrunch reports .
MC-Bench was created by Adi Singh and his team consists of volunteers. Leading AI companies, including Anthropic, Google, OpenAI, and Alibaba, provide subsidized use of their products for testing, although the website specifies that these companies do not participate in other projects.
Singh suggests that games can provide a safe and controlled environment for testing the “agentic thinking” capabilities of AI, surpassing the limitations of real-world testing.
A 21-year-old student created an AI application that helps programmer candidates deceive employers during interviews. Interview Coder has already been used by thousands of IT professionals. Tech giants are disappointed and ready to bring back in-person job interviews in offices
A 21-year-old student created an AI application that helps programmer candidates deceive employers during interviews. Interview Coder has already been used by thousands of IT professionals. Tech giants are disappointed and ready to bring back in-person job interviews in offices
American schoolboy becomes first person to win Tetris. Previously, only artificial intelligence could do it
American schoolboy becomes first person to win Tetris. Previously, only artificial intelligence could do it
Read the country's main IT news in our Telegram
Read the country's main IT news in our Telegram
