Yes, the event model is the only thing that makes sense. Just wanted to say, I don't think Kent's idea is in conflict with the Event Sourcing pattern, as long as these "Effective/Posting" events are stored append-only. Once you start updating the "Effective" date, you're losing data (a customer might call back and ask what the old Effective date was, to understand a paper bill, and you've lost it).
Event sourcing is a completely different way of thinking about business information. Of course you are going to run into things that look like "downsides". Consider the desired semantics of event sourcing and how your concerns run directly in the face of these semantics.
If you aren't comfortable with what qualitatively seem like "trivial" updates, then I don't think a large enough leap of faith has been made. We like to empathize with the machine but sometimes we take it a bit too far.
Event sourcing absolutely works and it works at the most massive scales the best. If you can dedicate an entire wing of your 15th floor office to "enterprise messaging and events", this stuff 5000% works. It seems to be at the small scale that it fails most frequently. At small scale, it's usually just one developer who has enough motivation to see it through, and the moment they shift their focus it falls apart.
I saw it work in manufacturing in ways I cannot do justice in HN comment size limits.