Why should anyone listen to Timnit Gebru or Emily Blender after how consistently they've been wrong about everything? What insight or value do they bring to this?
Not straight factually wrong but they try to give the impression that the models performance in maths is not a big deal but have you looked at the kind of maths they are actually doing? It's way above your average human probably in the top 0.1% or 0.01%
check out the Frontier Math Tier 4 examples <a href="https://epoch.ai/frontiermath/tiers-1-4/benchmark-problems" rel="nofollow">https://epoch.ai/frontiermath/tiers-1-4/benchmark-problems which GPT 6.1 just got 100% at. Up from 5% for Claude 4.5 last year say. And probably 0% for me although I was quite good at maths at uni.
square_usual · · focus · HN ↗
dgellow · · focus · HN ↗
tim333 · · focus · HN ↗
check out the Frontier Math Tier 4 examples <a href="https://epoch.ai/frontiermath/tiers-1-4/benchmark-problems" rel="nofollow">https://epoch.ai/frontiermath/tiers-1-4/benchmark-problems which GPT 6.1 just got 100% at. Up from 5% for Claude 4.5 last year say. And probably 0% for me although I was quite good at maths at uni.