One main obstacle is the complexity of the relationship between DNA and human traits. Will we ever know which genes influence intelligence and master how to manipulate them? However, as computer technology and understanding of genetic increases, it may only be a matter of time before a number of human gene-trait relations are defined clearly.
Another obstacle is the ethics of human experimentation. Failures in human experiments should not be accepted, so confirmation of the technology will not occur until animal GE experiments are considered extremely reliable. Some governments may choose to never allow human GE because the first child will always be an experiment with potentially unknown consequences. After animal experiments are reliable, will we move to GE on humans?? Fukuyama predicts that experiments WILL occur, especially with somatic gene therapy (for adults by their choice) if the risks of the disease outweigh the risks of the gene therapy. But for embryos, GE experimentation is a completely different matter. (He avoids predicting or arguing anything about embryo experimentation in this chapter--I wonder why? Does he deal with it later?)
Finally, Fukuyama talks about the obstacle or risk that GE will lead to altering human nature on a large scale. In other words, a whole population of humans may develop engineered traits and that may have consequences for human society such as discrimination against those who have or don't have those traits. If GE is only available for rich persons, that would limit the effect on the whole population, but if it becomes inexpensive, human society may be affected.
Mostly importantly, assuming that the technology becomes available, what are the ethical