We humans do a terrible job in seeing things for their constituents and the causes that led to the thing to appear. Doing marginally better in either of these fields could mean significant improvements in solving certain kinds of problems.
The more this problem of machine intelligence is looked at, the more it appears that “greater than human intelligence” is not only an oxymoron, but also not something we should seriously seek after.
Even it was somehow possible for a non-human to have human intelligence, it would automatically have all the cognitive flaws of humans as well.