I think the problem comes down to the HAM rules having been formulated pre-Internet, where the distinction between amateur and commercial activity was a lot more obvious: noncommercial ~= useless, commercial ~= useful. The advent of the Internet demonstrated the existence of a whole new category: noncommercial ~= useful.
Today we've got all kinds of noncommercial but highly useful RF applications, e.g. Meshtastic, that are stuck in the relatively limited ISM bands (yes, I'm aware of HAM mode, but recall usefulness → encryption). Just using the Internet isn't a replacement for these projects, as they're intended to be useful offgrid or during times of centralized Internet disruption. IMO it's a crying shame that stuff like this, the bleeding edge of non-commercial usefulness in RF, is lower priority than ragchewing. Besides, it's not as if all the technical aspects go away under looser rules: propagation behavior etc etc are just as important, if not moreso, than they ever were.
This doesn't mean just hand it over to the corps; there's a third way. Some chunks of current HAM spectrum should be converted to ISM. That doesn't mean a total free-for-all; duty cycle and EIRP limits apply and can be scaled appropriately to the frequency's propagation characteristics. The degree of freedom in usage in ISM bands drives innovation in protocol design to make better use of available bandwidth, as has happened with 802.11, so congestion/commercial use is actually a good thing.
Also, very specifically, 47 CFR § 97.113(4) states "messages encoded for the purpose of obscuring their meaning" are prohibited (with one exception in § 97.211(B)). The intent is to prevent illicit commercialization and promote self-policing of amateur spectrum, but it does not truly preclude using encryption, cryptography, authentication, and encodings so long as the "meaning" of the message isn't "obscured."
Also, AFAIK there are zero cases (i.e. FCC Notices of Apparent Liability) on intentional use of encodings for the purposes of obscuring the meaning of messages on amateur spectrum, and a very small number of cases of hams complaining about illicit transmissions, most of which were a result of an error or spurious emissions. I think most hams read the law and stray far away from brushing up against it by avoiding even the discussion of encryption carte blanche, but I think it's unexplored territory that would help to advance the state of the art, so long as we abide by common-sense practices to avoid *obscuring the meaning*, e.g. clear-text identification, public/published encryption protocols/keys, etc.
But if your intent is to basically use encrypted ham radios as an off-grid alternative to the internet and/or ISM/licensed wireless/cell networks for private or commercial use...then just use the internet and/or ISM/licensed wireless/cell networks.
It's also not far-fetched to amend the regulations. Debbie Lesko (R-AZ) is writing a bill to change the antiquated symbol rate limit to a bandwidth limit, opening up new modes, data rates, and capabilities for hams and emergency communicators on HF, VHF, and UHF spectrum. "Obscuring the meaning of the message" is very vague, and it certainly needs to be clarified and updated with the best interests of wireless experimentation at heart.