minus-squareVAK@lemmy.worldtoTechnology@lemmy.world•Announcing ARC-AGI-3 - A benchmark that tests if AI can explore, learn, and adapt in unfamiliar situations. Humans score 100%. Frontier AI scores 0.26%.linkfedilinkEnglisharrow-up2·3 days agoProbably AGI-42 linkfedilink
minus-squareVAK@lemmy.worldtoTechnology@lemmy.world•Announcing ARC-AGI-3 - A benchmark that tests if AI can explore, learn, and adapt in unfamiliar situations. Humans score 100%. Frontier AI scores 0.26%.linkfedilinkEnglisharrow-up0·3 days agoAI won them linkfedilink
Probably AGI-42