It is still a questionable proposition -- "10x engineers" are a thing, and brilliant singular insights that unlock billions of dollars in productivity like the concepts of convolutional neural networks and MapReduce are the brainchild of rare chart-defying individuals, rather than a product of the masses working in tandem. The Pareto Principle still seems to hold, even for software engineering (just ask any Bay Area software company's HR department).

Regardless, as an appeal to intuition, there is some sort of minimum bar (software engineers are still a VERY scarce commodity, as evidenced by market dynamics), suggesting that the average individual (100 IQ) is ill-equipped to take on the sort of abstract reasoning and novel thought processes so readily discarded by billions of years of prior evolution (indeed, knowing C++ would not have helped you as a primordial human fighting off that hungry sabertooth tiger) needed to comprehend and write robust code wading through some of the most complex systems (the Linux kernel is millions of lines of code) ever fathomed by humankind. This is the whole premise behind IQ theory, which really just boils down to a simple statistical argument.