2. There are no FDA approved therapeutics for disease that use bacteriophage, there is only some things as preservatives for meat. Most treatments of genetic disorders are generally nonspecific - they treat symptoms and gene therapy is still a developing field. Even in that case, the delivery vector is usually generalizable - usable in many patients with slight individual modification.
3. in vivo refers to in glass, or in an experimental set-up, this is most clearly not in vivo.
Anti-retrovirals only do so much - they have to be taken for the patient's whole life. This is different, they are claiming that HIV has effectively been eradicated in a living body, which just hasn't happened before. This isn't meant to be a clinical trial, it's the only living proof we have that we can kill all the HIV without killing the patient.
And you're mistaken about in vivo. In vivo = "within the living", in vitro = "within glass" (& "in silico" = "within silicon", ie. simulated on a computer).
If you do believe that organ transplants can cure a disease, then there must be some threshold of donor commonness that divides organ transplants and this treatment. Where might that line fall?
This case is a cure. HIV is pretty well known to not go away on its own (except under the circumstances described in your link). This treatment is almost certainly responsible for him no longer having HIV.
The notion that a cure must be widely applicable in order to really be a cure is rapidly becoming obsolete: many hypothesized or developing treatments involving stem cells or genetics or bacteriophages are very specific, often to the level of needing to be manufactured/cultured/etc. on a per-patient basis. This makes them less useful, but really has nothing to do with their ability to actually work.
This case is not announcement that a miracle drug will be hitting pharmacies in a few decades, but it is proof (or at least, very strong evidence) that HIV can be defeated in vivo.