minus-squarelath@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-up0·4 days agoBiased study. Take any average person off the streets and shove this thing in their face. That 100% notion will go down fast. linkfedilink
Biased study. Take any average person off the streets and shove this thing in their face. That 100% notion will go down fast.