It's just that I have seen projects delay making an important decision by rationalizing that there is not enough information and that the decision can be made later.

Then it turns out that the vital piece which fits the requirements does not exist at all (or does not exist for small time customer), or is incompatible with the architecture which has a lot of resources invested in.

Not saying that this will be what happens and most likely Librem team are already well aware of this risk, but it is certainly a possibility.

Yes, just on power evelope alone so many unknowns that could influence other parts of a ‘done’ design.

I wish them well but I’m sceptical. Still, it would be great to be surprised.

The problem is not knowing if it will work in my region and on my network. It's a rather big caveat for something with open preorders.

Based on playing with my little adafruit cell chip I have, I can talk to it pretty simply by sending it standard modem commands. I don't know how this exactly lines up with a more modern phone but it's pretty cool. You could theoretically talk to it with something like this too

https://geoffg.net/terminal.html

I hope they are able to create a free RIL for whatever modem they choose (in time).

That's actually not a big deal, as there are already multiple to choose from. FSO, oFono, ModemManager, and even more if you count ones from Openmoko times. Most of them should work with a module like PLS8 with minimal or no modifications.