Apple @lemmy.world

pulpier @lemmy.world

2mo ago

Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all. They just memorize patterns really well.