I've investigated many of the claims that one can teach intelligence. From Dual-N-Back to Head Start, they have all proven to have negligible effects on future outcomes and no effect on measured G. I think there are also pretty solid theoretical reasons to think that "learning how to learn" is not possible, as intelligence is that capacity that enables you to learn, you are strictly limited by your intelligence even when learning how to learn.
There is also increasing evidence in the literature that the theorized transfer effect (that learning one skill increases your capacity to learn other unrelated skills) does not exist.
These are not conclusions that I relish; I advocate them because I think they're true. Better a bitter truth than a sweet fiction.
However, I'm pretty sure I could find studies pointing to a raise in whatever intelligence metric for those with a college education (specially mathematics or physics heavy) versus a control group that went on to menial jobs.
Actually I'll just post the first few Google hits (only glanced by they severely contradict your claims):
1) http://ije.oxfordjournals.org/content/39/5/1362.full
Causal relation between education and intelligence metrics observed
2) http://onlinelibrary.wiley.com/doi/10.1111/j.1465-7295.2010....
Correlation between education attainment and IQ
---
Moreover, there's the even more glaring correlation (and causal relationship) between success and higher educational attainment, specially in technical fields, where obviously you need a lot of hard to learn knowledge, skills and general ability to learn (that is, I also dispute your claim that education has no effect on ability to learn). I can ellaborate and dig more if someone is interested/skeptical.
Interesting. Presuming that such a "transfer effect" is hypothesized to work the same way as transfer learning (in statistical learning, machine learning, cognitive science, etc.), this would be one of the first times mathematical models have predicted a cognitive capability should be commonplace that human beings failed to actually exhibit.
Maybe I misunderstood your point. Stories like that of Jaime Escalante suggest that there's a huge number of undiscovered talent and/or that better teaching makes more talent.
https://en.wikipedia.org/wiki/Jaime_Escalante