My 2 cents 
- Dual grounding scheme for balanced version: play vs calibrate. I did it and it works.
I think I know what you're getting at here but I would ask you to elaborate please.
- Shielding the analog path through LDRs from the rest of the board, which is digital.
To be clear the audio signal path in the existing design is both physically and optically isolated from the analog control path. However to your point the analog control does share a power supply with the digital and these will be separate supplies in the next design iteration.
- Improving the ergonomics of the Apple remote as it is overloaded of functions (use also the enconder + remote for setting impedance, display, calibration, etc.)
I agree and options include: a) Reorganizing certain control functions into a "maintenance mode" which has to be enabled otherwise it operates as a simple remote; b) Moving to a custom remote. This has its advantages but carries a hefty up front capital cost to get unit cost down. Plus the idea of yet another custom IR remote just seems so old school and annoying; c) Moving to a "soft" remote app for both iOS and Android. The biggest obstacle here is the lack of universal built-in IR in all smartphones but this could be done with a simple low-power wifi interface in lieu of IR interface or alternatively an intermediary wifi-to-ir device external to the product itself.
- Avoiding to use switching regulators: at least giving this as a paid option, eg using Belleson regulators.
I will disagree as to the idea that switching regulators are somehow inherently inferior to linear regulators in audio applications. The regulator we currently use is actually quite a nice bit of hardware and far superior to the ubiquitous LM9XXX series we used in the V1 model. That said, it's certainly not a super-regulator like the Belleson. Which by the way we are in the process of evaluating. Also, more recently there are some newer ultra low-noise linear regulators from TI and others that rival super-regulator performance. All of those are surface mount components. Also, anyone daring enough with a solder iron can swap out the existing regulator for a Belleson. Just get one with a LM78xx pin configuration and around 400-500 ma capacity. Be very careful removing the existing regulator which is rather beefy - easy to bugger the pads or lift a trace.
- Making firmware devoted to balanced version as it is different from the single ended version, eg impedance is doubled.
This may best be solved by multiplying by 2 in the interim but I get your point.
P.S. I am against using buffers, IMHO, as this alters the cool passive nature of Tortuga. A person sold his VTL 7.5 pre (yeah, that expensive VLT pre) as his balanced DAC was more transparent without, so using just the internal volume of the DAC. After trying my balanced DIY Tortuga, probably changed mind... he asked me if the Tortuga had a gain... of course no!
In my view there's a small contingent of early adopters who are open to the idea of using passives generally and within that small contingent there's a subset that may eventually discover Tortuga LDR passives and see the light!
However, I think the vast majority of audio enthusiasts are quite convinced that the only good preamp is an active preamp - evidence to the contrary notwithstanding. Thus a buffer option opens up the possibility of appealing to a larger market that may be willing the embrace the benefits of LDRs as long as it comes with the requisite familiar buffer/gain (i.e. it's an active preamp). Said differently, give the customer what he wants. Of course then there's Steve Jobs who said the customer doesn't know what he/she wants until you give it to them. Perhaps a buffer/gain bypass switch is in order.
