My start-up idea: “When your baby Julia (say) is born, you send off a cheek swab to Marvelton Labs, and they send you back an “owner’s manual” for Julia which will — most importantly — describe her cognitive tendencies so that you can “row in the same direction” with her as she grows up.”
When I’ve given this “elevator pitch” to adults who have raised at least two of their own children from birth, they immediately “get” the idea. Why? Because they have noticed that each of their children was different — their brains were different — from birth. But without an owner’s manual, they have struggled to figure out how to best support the different strengths and weaknesses of their kids.
“Sally” cruises through school, learning effortlessly, working hard, and getting straight A grades. “Bobby” is bored in school, doesn’t do his homework, and acts out. He does not put in the time he needs to on homework, and his grades suffer quite a bit. Bobby’s parents move him around to different public and private schools, but nothing seems to make a difference.
There are several flavors of IQ tests, the more exhaustive (and exhausting) Woodcock-Johnson Tests, and several other test “instruments” that attempt to assess cognitive function in human beings. But these tests are expensive ($3K in Seattle in 2013) and time consuming (2 days) to administer, so few children are tested, and usually only when they are really struggling in school or at home.
The way to implement this idea is conceptually quite simple: a) gather detailed cognitive performance data on (say) 1M teenagers and adults, b) collect DNA samples from those same 1M people, and c) run a big machine learning exercise to find patterns.
Most parents I have talked to would love to have (had) this for their children. A few express concern about “predetermination” or “ruining the surprise”. Obviously there are very important privacy implications for this technology.
Ben Slivka 
One thought