Yeah, this isn't remotely true. It's probably the most reliable trait in all of behavioural science.
Also, given the fact that you can study for IQ tests, I don't really think I agree at all that it's some immutable and objective fact of behavioral science.
And yes you can study for IQ tests, but I could study basketball and while I might get better, I'm never playing in the NBA.
I'm not saying that IQ is useless or anything, just that we should take the results with an appropriately large grain of salt.
What you are seeing is that if you cheat it helps a little bit. That's not an argument for OR against them in general, except maybe to say that they should be more resistant to this type of cheating.
About half of general intelligence is this correlation between all the different types. Modern IQ tests measure something like 15 types, and pattern recognition is one of them. It's the correlation between all the types that really matters.
It is strongly linked to career success, educational attainment, and many other things in life that are typically regarded as positive outcomes. Lower g-factors correlate strongly with poor outcomes.
Regardless of what we're measuring (and if it can be improved with various interventions) it is something we can measure, seems to be heritable, and is strongly correlated with positive outcomes.
The GP's phrasing was harsh, but he's not wrong for insisting that good and bad are not the same thing.
Below average intelligence is something that should be fixed, not celebrated, and it's not just some cute and quirky Sesame Street "We're all different, but the same" thing. It's a disorder, like Depression or Schizophrenia. It should be treated, and ideally eradicated. People who have it should not be stigmatized for it, but they SHOULD be treated and at least taught coping skills - as they would be with any other disability.
That said, even in the article you linked it does not appear to be a universally accepted truth.