> Either way, how the hell do you get a patent for some abstract diagrams that represent a design pattern.

You don't.

Patents are unfortunately very obtuse to read and hard to parse, and it's especially easy for a lay person to give up very quickly and as a result misinterpret the patent. This is especially true when you're pointed to the patent by a rando on the internet telling you what (they think) is patented, even if it's not accurate.

Another contributing factor is that, even when you read the claims, a lot of the actual description may be functionally irrelevant to the actual thing being claimed. Suppose, for a moment, that you were patenting a new kind of bike pedal. Then your actual claim would look like an overly verbose and vague description of a bike that includes your new kind of pedal. And since most of the paragraph is describing a bike, it's easy to assume that what is being patented is the bike and not the pedal of that bike.