It's great to see Stacken's old machine being taken care of and tinkered with, I'm sure some of our older members will be thrilled to see this. For the past 20 years we sadly haven't had space for minicomputers or mainframes, though for a while we were planning to run an emulated PDP-10 with ITS or something (project on hold due both to the emulated system's clock drifting dramatically and the VT220 I was using spontaneously breaking very subtly: key commands don't transmit anything, but everything else works). I'm too young to have used any of these machines myself for anything approaching real work, but I can't help but be fascinated by them.
A pair of smart modems should certainly be able to connect, but it's very likely they would only be able to negotiate a fairly low speed. The classic 300 baud Bell 103 modulation can probably work perfectly fine even over early 2010s Skype. Although possibly with a lot of dropped bits...
I remember dialing into a BBS in the mid 2010s over an international phone call going through a late-stage digital switch locally (possibly one of the later AXE switches but I have no way of knowing for sure anymore), and the call would almost have to have gone through lossy compression on its way to the US.
As far as I know calling free is optional with Fil-C. It can prevent situations where some memory is no longer needed but isn't freed due to some remaining pointer to it sitting around somewhere, but it's not required. The GC will act like a GC if given the opportunity.
Typically, for development, one will start a clean image in the morning and load it up with a combination of newly compiled code and automatically cached previously compiled code. This is mostly invisible to the user unless something goes wrong with the cache and it has to be cleared. The unaware user could think they're recompiling everything every time.
