That's one of the worst I've seen, that 1 MOhm resistor does nothing at all. I've seen this variations with a couple hundred ohms, which at least could plausibly make sense in some situation. But when in doubt just connect all the grounds together. And yes, the ESD diodes need to connect to the same ground as the chip or they won't work correctly, and can potentially conduct ESD to your chip.
There are some really good application notes out there, but a lot of them are written by people 2 years out of university so they know the fundamentals but have very little practical experience yet. They have gotten some incomplete guidance from more senior engineers (who might also not really understand the issue, but may have better developed "instincts") and they search around on the internet or other datasheets, or their companies internal designs and try to make up a good story about what they see. So often what you get is very similar to the output of a game of telephone or a large language model.
So if you are reading an application note or application section of a datasheet if something doesn't make sense, don't assume the problem is just you. The AN could be completely wrong. It could be right but the explanation is wrong, or it could be right but only in a situation completely different from what you are interested in.
Textbooks are typically better vetted than application notes, since they are more often written by experts who have both training and experience. But they are updated less frequently and usually written to be more generic, so you may need to do some more work to adapt the information to a particular situation.