Not sure what you mean: we can sign TI-68k FlashApps with TI's private key, and they're accepted by pristine OS versions.

I meant that we don't have the private key. Or do we now?

I think another issue is how it's possible to run programs that are archived on 68K calcs, which is impossible on Z80 calcs, making it even more pointless to make Flash APPs on 68K (since one of the main reason why they're used in the first place is to save RAM)

As I wrote in the other topic, the size limit has been meaningless for practical purposes nearly since its inception, in the 20th century
There are several multi-chunk ASM programs which use 100+ KB of code in RAM.
GTC is a FlashApp because having the code execute in RAM would have severely reduced the amount of RAM available for storing compilation data and/or the speed. For functionality and optimization reasons, it's mostly compiled with TIGCC/GCC4TI to COFF object files, the toolchain in TIFS is used mainly for the final FlashApp production.

OK, I'll try to get the IDE to be a flash app, then

Ouch, it's a bad idea to add a TIFS dependency to your program...
While FlashApp signing is now trivial, and GTC shows that it's possible to reduce the pain of using TIFS, it doesn't mean that FlashApps are a good thing to do, especially in this day and age...

since I don't know much about using multi-chunk ASM programs

To begin with, it's easier than making a mixed GCC4TI / TIFS compilation, especially if you use anything newer than Windows XP. The TIFS IDE works only on XP and earlier, due to a dependency on Microsoft's proprietary, non-standard, "Microsoft Java" (and it's horribly buggy anyway); the toolchain may work on newer Windows versions, but then you'll have to invoke it from the command line.

the RAM needs to be used for the TI-Basic programs to edit.

That's a fact, but there's room before your editor becomes so big that it can't fit in RAM alongside the BASIC program...