The devil is in the details.
For example, it could be that the phy and link layers are interoperable but the network layer is not (and hence by definition no higher layer is in practice interoperable) - so you can exchange packets with a single adjacent node, regardless of vendor, but not either relay (route) or actually do anything with the packet.
Or it might not even be as good as that, for example, because there are too many protocol options even at the lower layers.
I note also that the goTenna web site talks about options that are only available with a subscription and intentionally limited functionality otherwise. Hmmm.
There is always tension between a competitive advantage in the market and encouraging wider adoption. My opinion, for what it is worth, is that if Semtech wants this to be a success then they need to drive total interoperability even if it doesn’t suit the agenda of manufacturers using Semtech’s chips or patents. That could mean having formal interoperability certification.
In some respects it reminds me of the internet before there was the internet.
I suppose the military may want their own frequency but it is not as if this would prevent jamming, and the military already have the option of having their own frequencies.
One potential use of this is in disaster areas where all terrestrial telecommunications infrastructure is dead. In that case, it would be more helpful if everyone actually can communicate with everyone else, including the military.
Unclear how a “professional type” might differ from a “mere mortal type”.