Unicorn, pelican, Middle-earth: OpenAI co-founder Karpathy is looking for the next AI vibe test

One paragraph of "Lord of the Rings" in, 5,500 lines of code out. Andrej Karpathy had Claude Opus 5 turn Tolkien's opening into a 3D browser scene. The article Unicorn, pelican, Middle-earth: OpenAI co-founder Karpathy is looking for the next AI vibe test appeared first on The Decoder .

The top 3

  1. The 'Pelican on a Bicycle' Challenge: This benchmark required AI models to generate an absurd SVG image of a pelican riding a bicycle, testing spatial reasoning and creative interpretation.
  2. ImageNet's Foundational Visual Recognition: ImageNet was a cornerstone benchmark from 2009 to 2017, challenging AI systems to classify photographs into 1,000 categories and driving significant advances in deep learning for computer vision.
  3. Karpathy's Middle-earth 3D Scene Generation: Andrej Karpathy's recent experiment with Claude Opus 5 involved generating a 3D interactive web environment from a Lord of the Rings paragraph, demonstrating AI's advanced code generation for complex worlds.

Sources

Open the full topic