I got the chance to meet Patrick Chung at Stanford once. If he says that's what 23andme can become, I honestly believe it. He showed incredible insight into startups and emerging markets when I met him.
In my opinion (as a statistician with minor bio experience) the future of predicting health things based on just DNA is uncertain, except for direct remedies for known current conditions like certain forms of cancer. (i.e. as a response to disease, in actual personalized medicine)