A new, challenging AGI test stumps most AI models
SMRTR summary
The Arc Prize Foundation's new ARC-AGI-2 test challenges AI models with visual pattern puzzles, revealing a significant performance gap between AI (scoring around 1%) and humans (averaging 60%), while aiming to measure AI's ability to learn new skills beyond its training.
SMRTR provides this summary for quick context. The original article belongs to TechCrunch.
Read the original article