There's no "most optimized human". We are already that, perfected in millions of years. What could really happen is the split between multiple sub-species. For example, it makes perfect sense to do the optimization for orbital station dwellers or Mars colonists or underwater dwellers.

I also wonder what happens if this kid one day has kids. In this case it was a very rare genetic disease, but if the same was applied to a less rare genetic disease (where it is also more beneficial to have a treatment as more people have use of it) wouldn't the end result be that more and more kids will be born with these diseases?

Can it be applied to adults? Useless for this particular disorder, but what about others?

Low hanging fruit is very low hanging in this case. There are many point mutations for example that confer risk to disease and cancer. Lynch syndrome which confers significant risk for colorectal cancer for example is something that could he cured with transgenic humans today even with todays technology. Just a matter of screening gametes for the mutation (usually one base in the case of Lynch in heterozygous state with wild type healthy allele and that wild type healthy allele gets a second hit mutation as the cancer develops and things just go off the rails from there) and editing that base back to wildtype. No downside only upside with that.

What gets harder are polygenic traits that even today we don’t have great data on what are the causal alleles. But that is also not a technological limitation either but a statistical one from insufficient sampling of these polygenic phenotypes.