Andrew Ng: We need harder evals — have frontier models chat with me

andrewgwils · x · 2026-09-04

Andrew Ng posted that current benchmarks may no longer be sufficient, jokingly challenging people to have their frontier models chat with him so he can ask questions that would "make their heads explode." The underlying point: frontier models still struggle with sufficiently hard open-ended questions, and the community needs harder evals.

Original post →

More from Models

Models channel →