Thanks everybody for input, it’s really appreciated and useful. I’ve been tied up with Actual Work so have not moved anything forward in the past week and a bit - hence it was a good time to let this thread grow.
I am working on deliberately not making this:

and so its unlikely functionality will change much; there will certainly not be any more significant features added.
The relatively naive CV out solution (buffered via the faders direct) is designed as “just enough for a surprising number of use cases, and much better than nothing”. And if you don’t want to use the sixteen jacks along the back: nobody is making you use them! It still works fine as a MIDI box (and, for the people who’ve mentioned devices like the FH-1, for instance, it would work fine patched straight into that). I would also not make judgments about the final hardware based on images you’ve seen of prior versions - I know there’s been talk amongst ourselves of aluminium top-plates, for instance, which change the weighting a lot. But similarly, don’t imagine it will change hugely significantly beyond materials used and minor dimensional changes. I can confirm, form-factor wise, it is still the width of a 128 grid, and still slightly shorter. I’ve always been using the 128 as a constraint, size-wise.
There’s nothing to stop people exploring building external expanders with added functionality (such as controllable ranges/curves, wider ranges, yadda yadda); but I am not designing and making two complex electronic things at once. Making two things at once is no fun at all, and is more than twice as hard as making one. Making one, in our spare time, is hard enough.
Don’t forget, if you have opinions of your own about this design, you are welcome (when they’re published) to all the open source files and to make your own modifications/forks.
So: next steps for us are probably finalising a few choices to get prototypes built, and we’ll see how that goes.