I like Plone 5. I'm porting a Plone 4 application to Plone 5. I've been through upgrading existing applications from every major version since Plone 1. I think the degree of change for display technology is higher on this front than with previous major versions, in part because we got all the following at once:
- AMD (good for core, bad for add-ons)
- Patterns (good, generally, IMHO)
- New server-side registry (confusing, but getting better).
- Old ways of configuring actions need changing (icon story still too complex).
AMD forced the new registry; the new registry's tight coupling of CSS to JS (with particular constraints) added another wrinkle. We can mostly deal with the some of the biggest warts of the registry now that stub_js_modules functionality is in 5.0.x, and bundles will be concatenated into single download in 5.1 (meta bundles PLIP).
Plone has had a "JBOA" (just a bunch of add-ons) integration ideology for a long time, and the new JS story was not very considerate of making that ideological fit easy anymore; we can do this in spite of AMD (and with more work), not because of it. All, of course, may be moot in 3-4 years once all browsers support ES6 imports, but I don't think it an unreasonable opinion to have a love/hate relationship with Plone 5's JS/CSS story.
My frustrations, for the record, are not holding back my work on upgrading my own applications to Plone 5, but I do have to recognize that there's slow progress to weigh against clear UX benefits motivating me to do so in the first place.