VAK

joined 1 week ago
sorted by: new top controversial old
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%. in c/technology@lemmy.world
[โ€“] VAK@lemmy.world 5 points 3 days ago

Probably AGI-42

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%. in c/technology@lemmy.world
[โ€“] VAK@lemmy.world 3 points 4 days ago (3 children)

AI won them