A new, challenging AGI test stumps most AI models

Comments