No, the FDA thinks it can decide whether someone can sell a product claiming to diagnose disease. Because, fairly explicitly in the Food, Drug, and Cosmetics Act, it can, and, in fact, is obligated to.
23 and Me gives you the raw genetic report. If you're sophisticated enough to not panic and jump out a window because you have some terrible disease, the assumption is you can also find open source data and/or software that will give you this same information.
The problem with what 23andme was doing is going direct-to-consumer with tests that were potentially sketchy. If you're willing to risk sketchy information you can find all kinds of bleeding edge research on your particular genetic makeup and choose how to handle it.
Adults can't handle misleading and inaccurate health information, and are known to spend large amounts of money to e.g. literal snake oil peddlers back when peddling snake oil as a cure-all wasn't prohibited.
Yes, we are regulating who's allowed to tell people that a relative is going to die, and we're asking people who do so to show evidence that they know what they are talking about. If someone would go around selling a service "is your relative going to die" by guessing or simply telling what they want to hear, then that should be regulated and prohibited.
As another poster said, "One problem is that they warn that your offspring are at high risk for some condition, when really "high risk" means 0.5% higher risk than the general population. The other is that they may say you are not a carrier for a certain condition, when they only test for one variant of it, where proper tests will test for multiple variants."
If you tell people "we ran a test for X and it was positive/negative" then you'd better be able to show that whatever rituals you performed actually lead to reasonable information about X. Simply having a test that has some information related to X (e.g. if it would be used together with other factors to diagnose X or not X) doesn't mean that you can honestly describe it as "test indicating a high risk of X" - it may be that this particular test is indicating that, and it may be that it (alone) is misleading, and we need someone (e.g. FDA) to draw a line.