I keep the -1 from my review for the original proposal. The revised proposal is still focused on JSON exclusively and ignores the fact that Codable is supposed to work for other serialization formats.
As other reviewers have pointed out multiple times, there's no sane default here that can work well for all (or at least the majority of those widely available, like XML) formats. The inability to sufficiently customize it is a deal-breaker in my opinion. I don't see that much has changed from this perspective when comparing it to the original proposal.