Plenty of websites are implemented poorly, but 3/4 of the problems listed with Flash were inherent. Flash absolutely was a problem.
It did have a certain (imo over hyped) DX that people had a fondness for & that the web wasn't catering to. This was good for users in that they got a lot of experiences they might not have otherwise had. But the experiences were 99% things the web could do, and the web experience as a user was better.
I'm not unsympathetic to the passing of the flash age (even given the treasonous un-webful proprietary closed app-like rot that it was). But I mostly wish we'd had more people writing competitive authoring tools for the web; we could have made a Flash timeline like editor, with svg and smil/svg Dom/css animations/css transitions. The DX was largely recreateable. I think it's somewhat over hyped, but I definitely see & am sorry that there haven't been successful popular web-targetting ways of doing similar.
I don't think the web is as impermanent as you project. Many sites will run fine from an archive. Sure, datavis sites not being able to get data you didn't fetch the first time is an issue, but flash dataviz's are just as liable to fetch data on the fly rather than bake it all in. Flash doesn't have an inherent advantage, more a happenstance one, and web archiving will take one far further than one might believe. Alas "save page as..." doesn't work; the promise of tech like http bundled seems high, but then the archival community groups went & did their own thing instead of joining up, and http bundles became another bland anti-Google rallying flag. Alas.