X debate challenges view of LLMs as databases that merely rearrange text, citing frontier models' near-perfect SAT and GRE scores plus PhD-level research outputs as evidence of distinct capabilities
— One response frames LLMs as neural networks sampling from learned distributions.
——0——
6676114.8K
Cluster engagement
10 snapshots
Sentiment
Pos0%
Neg100%
Users are dismissing debates on whether current AI models match human intelligence as pseudoscientific or overly complex while using personal insults against other participants.