I'm a different person weighing in here:

When you said:

The T3SS is one of the most complex bacterial molecular machines, incorporating one to over a hundred copies of more than 15 different proteins into a multi-MDa transmembrane complex (Table 1). The system, especially the flagellum, has, therefore often been quoted as an example for “irreducible complexity,” based on the argument that the evolution of such a complex system with no beneficial intermediates would be exceedingly unlikely. However, it is now clear that, far from having evolved as independent entities, many secretion systems share components between each other and with other cellular machineries (Egelman, 2010; Pallen and Gophna, 2007).

I ofc am just a layman reading this, I agree it seems better understood that how I interpreted what he was saying, but it also doesn't seem nearly as well understood as you're saying.

IMO it's a problem with the article. The article says that T3SS is cited as an example as something that's "irreducibly complex". I suppose that it's true that it is cited as that. But the second part of the paragraph explains why it isn't true that it's "irreducibly complex". The paragraph isn't explicit enough because the paragraph has probably evolved to be something that's true and equally dissatisfying to both sides.