Lots of "or" in that list that give them cop outs between what's right and what's cheep.
Remove or redesign all 16,000 barriers
Indeed. I accept that some barriers exist for legitimate purposes (controlling livestock, or keeping cars off the path) and might reasonably be replaced by a more accessible gate/cattle-grid design or the now legendary single bollard, that's a lot of weasel-room if they want to take it.
Transform the Network by replacing existing on-road sections with new traffic-free paths or by creating quiet-way sections
And there's nothing in that list to say "make routes that go directly between places people want to go"
This is the big one, I think. While they do commit to a traffic level for their definition of 'quiet', any traffic at all is a barrier to some path users (particularly independent children), and a strong disincentive to less confident cyclists.
And there's the fundamental lack of commitment to whether they're building a transport network (as the name suggests) or a series of leisure routes (what the NCN has actually achieved). Personally, I think they'd be better off accepting that they're doing the latter. Proper cycle paths for transport are - for the most part - going to be boring, direct stretches of tarmac/concrete alongside main roads, and that's surely the domain of local councils, rather than a railway-path-reclaiming charity. The off-road paths that are actually useful transport routes in urban areas are treated as useful paths by local users, not part of some network that takes you from A to B via C, D, F and J.
There are some lovely on-road sections of NCN in rural areas, but since The National Byway is a thing, why obfuscate the issue by including them? And us cycle-tourist types can see the yellow lines on OS maps just fine without little blue signs (though I accept they may function to publicise the network).
Now if they did something bold like "we will remove NCN designation from any path not meeting our new standards" I might have hope. For now, I doubt much will change.
That's point 9 in the list. I suspect the sticking point is the new standard will be too low.