You are viewing a single comment's thread from:

RE: LeoThread 2025-02-09 12:40

in LeoFinance4 months ago

Part 3/10:

Next, we see the trajectory of performance on Google’s proof question-answering benchmark. This graph, spanning back to July 2023, illustrates a dramatic leap from random guessing accuracy—25%—to a staggering human expert-level performance exceeding 70%. This rapid climb raises valid questions about the timeline for AIs surpassing human capabilities across a variety of tasks, from coding to expert-level domain knowledge.

Misjudging AGI Predictions