For every gene we could search for the best one that's out there and exchange the human version with it.
This would be an extraordinarily bad idea especially from yeast in particular. 'Best' is contextual.
It's contextual but that doesn't mean that "best" doesn't exist.
At the beginning it's unlikely to be tried on humans, but when it comes to farm animals I would expect that people do experiment with it. Maybe a pig with yeast mitrochondria does better than a regular pig because the yeast mitrochondria had more time to be highly optimized by evolution. It might be that you need to do a few additional changes to have the pig deal with the yeast mitochondria, but people will experiment.
If you get an enzyme that's twice as efficient in catalyzing some reaction, you have to down regulate it's expression.
I previously wrote a post hypothesizing that inter-group conflict is more common when most humans belong to readily identifiable, discrete factions.
This seems relevant to the recent human gene editing advance. Full human gene editing capability probably won't come soon, but this got me thinking anyway. Consider the following two scenarios:
1. Designer babies become socially acceptable and widespread some time in the near future. Because our knowledge of the human genome is still maturing, they initially aren't that much different than regular humans. As our knowledge matures, they get better and better. Fortunately, there's a large population of "semi-enhanced" humans from the early days of designer babies to keep the peace between the "fully enhanced" and "not at all enhanced" factions.
2. Designer babies are considered socially unacceptable in many parts of the world. Meanwhile, the technology needed to produce them continues to advance. At a certain point people start having them anyway. By this point the technology has advanced to the point where designer babies clearly outclass regular babies at everything, and there's a schism between "fully enhanced" and "not at all enhanced" humans.
Of course, there's another scenario where designer babies just never become widespread. But that seems like an unstable equilibrium given the 100+ sovereign countries in the world, each with their own set of laws, and the desire of parents everywhere to give birth to the best kids possible.
We already see tons of drama related to the current inequalities between individuals, especially inequality that's allegedly genetic in origin. Designer babies might shape up to be the greatest internet flame war of this century. This flame war could spill over in to real world violence. But since one of the parties has not arrived to the flame war yet, maybe we can prepare.
One way to prepare might be differential technological development. In particular, maybe it's possible to decrease the cost of gene editing/selection technologies while retarding advances in our knowledge of which genes contribute to intelligence. This could allow designer baby technology to become socially acceptable and widespread before "fully enhanced" humans were possible. Just as with emulations, a slow societal transition seems preferable to a fast one.
Other ideas (edit: speculative!): extend the benefits of designer babies to everyone for free regardless of their social class. Push for mandatory birth control technology so unwanted and therefore unenhanced babies are no longer a thing. (Imagine how lousy it would be to be born as an unwanted child in a world where everyone was enhanced except you.) Require designer babies to possess genes for compassion, benevolence, and reflectiveness by law, and try to discover those genes before we discover genes for intelligence. (Edit: leaning towards reflectiveness being the most important of these.) (Researching the genetic basis of psychopathy to prevent enhanced psychopaths also seems like a good idea... although I guess this would also create the knowledge necessary to deliberately create psychopaths?) Regulate the modification of genes like height if game theory suggests allowing arbitrary modifications to them would be a bad idea.
I don't know very much about the details of these technologies, and I'm open to radically revising my views if I'm missing something important. Please tell me if there's anything I got wrong in the comments.