And like all types of performance test the "something" it measures is a vector of pattern matching ability (both the component generalisable to solving more difficult problems and the component that doesn't except in the domain of career-relevant tests) but also other things obviously correlated with future performance but not fitting most people's definition of intelligence (attitude towards test, comfort in guessing, time management in timed tests, and to a degree experience), so whilst we know what the scale of the apparent impact is, we don't know really know what's behind the cause
We knew of gravity much before we knew of electricity. Try explain electricity to somebody in year 1300 and you'd be burned as a witch.
And yet we understood electricity much earlier than gravity. How electricity works is much simpler to explain than gravity.
— Alfred Binet, inventor of the IQ test
The original author has a good point about the naming of the big 5 factors though.
Two things I take issue with though - IQ is reasonably well understood in that we know it measures something reliably and accurately. Exactly what that thing is though, is not really understood - and probably can't be. And how it relates to the real world is also somewhat poorly understood, and may not be possible to fully understand. But it does have good ecological and construct validity. I'm talking about good IQ tests here - any less validation than the huge amount that the Wescheler scales go through and the validation is questionable. I'm aware of the racial biases in IQ testing which is a significant problem.
Psychology isn't a particular young science - big chunks of modern statistical analysis have their roots in 19th and early 20th century psychology. So psychology is a young science in the same sense that modern statistics is. Back when I was studying psychology 20 to 30 years ago, a lot of what is now called "neuroscience" was then called "cognitive science" or more rarely "cognitive neuroscience". Neuroscience sounds sexier though.
I do have a massive issue how many social scientists abuse psychometric techniques in inappropriate ways. Management research is particularly guilty here, and I've been pressured into doing this kind of inappropriate analysis in the past.