"It’s worth noting that most working professionals do a lot more than submit research reports to their boss, which is all that GDPval-v0 tests for."
"OpenAI says that it believes Claude scored so high because of its tendency to make pleasing graphics, rather than sheer performance."
So by wide range of jobs, they mean jobs consisting of only submitting research reports with pleasing graphics. And based on recent measurements, about 1/4 incorrect or entirely hallucinated.
Exactly. I was coming to say that.
Even so in the parts of Alaska outside of the Arctic Circle there's summers where the sun rises at 3AM and sets at 9PM, then in the winter it's sunrise at 9AM with sunset at 3PM.
In Fairbanks at the winter solstice the sun rises at 10:58 and sets at about 14:40.
At the summer solstice it's 02:57 and 00:47. Even though the sun officially sets year round it never really gets dark between mid April and mid August because the night periods never leave the twilight stage.
That *should* mean there are more males at the bottom end where people really need help
More at the top and more at the bottom, and the effects on earnings are not symmetrical. Once somebody is cognitively deficient enough to be considered a ward of the state the amount of resources they require to survive remains more or less constant with diminishing IQ but on the other end of the curve income potential does not have an corresponding cap.
but there are no such differences in average IQ between males and females
I've seen claims to the contrary stating that there is a small difference in the average and a significant difference in standard deviation.
The perversity of nature is nowhere better demonstrated by the fact that, when exposed to the same atmosphere, bread becomes hard while crackers become soft.