The Meme Machine by Susan Blackmore.
I remember a sort of consensus from the 00s that memetics had failed as a research program and the big-name people like Dawkins and Blackmore moving on to other stuff. Here's one summary I found. People still find the metaphor compelling, so it might just be that right now nobody has a good idea how to study the thing rigorously.
In certain circles, it's common to talk about memes as entities, in a metaphorical and handwaving way. But it seems to me that thinking rigorously and precisely about memes in the framework of "memes::genes, memeplexes::infectious agents, minds::hosts" actually has a lot of explanatory and predictive power. This kind of thing ought to be a tool in the rationalist toolkit. In the space of politics, for example, it can be very illuminating to strip away the content of an idea and analyze it purely as a replicator, or as a component of a larger replicator conferring a specific fitness advantage.
A cursory online search for existing literature on memetics reveals that the field is about as confused as you would expect. But perhaps I am going about my search in the wrong way. Is there another field that encompasses the kind of thing I am gesturing at here?