Let me make a simple parsimonious assertion: knowledge acquisition is limited to the precise information acquired with zero secondary benefits to any other areas of the brain. Furthermore, unless practiced regularly this information will most likely not be retained.
While such an assertion is in all likelihood not true, in the absence of evidence, it is more likely to be true than a more complex theory of how knowledge is acquired and retained.
Now, if I was to put Paul Graham’s argument into precise, scientific terms, it would be:
Holding time constant, time spent on more difficult problems is more likely to produce measurable improvements in fluid intelligence than time spent on simpler problems.
Or possibly: holding time constant, time spent on more difficult problems is more likely to produce measurable improvements in conscientiousness than time spent on simpler problems.
I’ll stick with the first definition for simplicity.
But increasing fluid intelligence is a hard thing to do. This famous study argues that solving working memory tasks increases fluid intelligence however this meta-analysis argues that the evidence in this area is still inconsistent. Even if working memory tasks do generate increases there is still the problem of whether it’s better to solve a few really difficult tasks or lots of easy tasks or somewhere in between. Then there is the issue of how long this increase is retained. Is it something like exercise where benefits mostly disappear within 6 months? And even if working memory tasks do increase fluid intelligence, that doesn’t mean difficult math problems increase fluid intelligence.
My own impression is that the best way to increase intelligence is by just reading a lot. As evidence, this study found that reading was the best predictor of increased success on a standardized test. The next highest predictor was being social. Studying had a positive, but non-significant impact. This was an observational study so the usual caveat of correlation may not be causation applies.
I believe that reading more complex literature up to a certain point is probably better than simpler literature, but I don’t have evidence of this. I’m also not sure of what that certain point is where complexity is too great.
Simple: Where’s the evidence?
Let me make a simple parsimonious assertion: knowledge acquisition is limited to the precise information acquired with zero secondary benefits to any other areas of the brain. Furthermore, unless practiced regularly this information will most likely not be retained.
While such an assertion is in all likelihood not true, in the absence of evidence, it is more likely to be true than a more complex theory of how knowledge is acquired and retained.
Now, if I was to put Paul Graham’s argument into precise, scientific terms, it would be:
Holding time constant, time spent on more difficult problems is more likely to produce measurable improvements in fluid intelligence than time spent on simpler problems.
Or possibly: holding time constant, time spent on more difficult problems is more likely to produce measurable improvements in conscientiousness than time spent on simpler problems.
I’ll stick with the first definition for simplicity.
But increasing fluid intelligence is a hard thing to do. This famous study argues that solving working memory tasks increases fluid intelligence however this meta-analysis argues that the evidence in this area is still inconsistent. Even if working memory tasks do generate increases there is still the problem of whether it’s better to solve a few really difficult tasks or lots of easy tasks or somewhere in between. Then there is the issue of how long this increase is retained. Is it something like exercise where benefits mostly disappear within 6 months? And even if working memory tasks do increase fluid intelligence, that doesn’t mean difficult math problems increase fluid intelligence.
My own impression is that the best way to increase intelligence is by just reading a lot. As evidence, this study found that reading was the best predictor of increased success on a standardized test. The next highest predictor was being social. Studying had a positive, but non-significant impact. This was an observational study so the usual caveat of correlation may not be causation applies.
I believe that reading more complex literature up to a certain point is probably better than simpler literature, but I don’t have evidence of this. I’m also not sure of what that certain point is where complexity is too great.