MER guided placement is still the gold standard, and image-guided approaches are primarily used for efficiency, not for quality. The factory-style surgical groups that push image-guided placement do so based on some very underpowered studies showing non-inferiority of outcomes and anatomical placement at the group level. However, as anyone who has been in the OR a lot can tell you, there are plenty of times where the anatomical placement of an electrode looks great, but MER and intraoperative testing reveals it to be suboptimal. Subsequent adjustments sometimes look less accurate from a pure imaging standpoint, but produce better functional effects. A pretty decent proportion of patients would receive leads that look great on post-op imaging but which are not ideal from an electrophysiology standpoint with asleep DBS (i.e.
https://www.sciencedirect.com/science/article/abs/pii/S187887501932248X).
That's not to say that asleep DBS doesn't have its place. A lot of generalized dystonia patients could never handle an awake case, and there are a decent number of PD and ET patients whose anxiety would make an awake OR situation... unpredictable. In those cases, the option to use an asleep procedure, with the informed consent that there is a risk that the placement may not be quite as "individualized" to their neurophysiology, is a nice thing to have, especially with modern directional DBS to tweak the effective placement in programming clinic. However, I would never send my family member to be treated at a DBS center that is only doing asleep cases. Not only does this leave out the possibility of a better placement with MER and exam guidance, it speaks to the values and approach of the center that's made that decision.
As far as programming, I think most high volume DBS centers are integrating APPs into the programming in some sense. Where I trained, programming was done mostly by nurses (not even NPs until after my fellowship) who had literally decades of experience programming many cases per day. Where I am now, I do more of the initial programming and more difficult cases, but APPs definitely offload at least half of our programming followups. I find it necessary to give those with less than 10 years of experience a clear plan to follow, but it's really not hard to bump amplitude by 0.1-0.2 mA in response to an increase in tremor. In the ET patients in particular, however, it's often necessary to keep a close eye as distinguishing ataxia from tremor can be difficult for a non-neurologist.
I really have less optimism about closed-loop DBS than most people. The extreme focus on beta-band optimization in these trials seems to work in some patients and not others. I saw several posters at MDS this year from Japanese groups that are already using closed-loop systems in clinic, and the takeaway was that you really can't rely on it to do your job for you, as in a pretty decent portion of patients (IIRC 20-30%) it successfully suppressed the beta band, but the patient was left woefully undertreated until manual increases were implemented. I don't know if this is because we need a more comprehensive understanding of PD electrophysiology, or because there's a variability in this EP signature based on implantation error, or because the entire concept of recording only from the same place we are stimulating is fundamentally flawed. Probably all 3 to some extent. My sense is that once we get it running, it will act more like the extra patient parameters we give to allow patients to turn their amplitude up or down a few clicks - not a replacement for experienced DBS programmers.