To be clear, the difference between Waterfall (as Royce described it) and Agile (as it was intended in the early days) is the iteration length (and regularity, but it's hard to have short, uneven iterations). The idea was that shorter iterations would have significant and beneficial effects on behavior. I wish I'd got to live in the universe where a couple of orders of magnitude more people had had the chance to experience that Agile. It is really nice.
So much of this is just wrong, not reflecting what actually happened.
And I'm surprised that welcoming changes late in development is commercially unworkable, doing it as I was in 2000 and 2001 at IBM. 🤷
It broadly gets one thing right: Agile as a reaction to the failures of single-pass Waterfall (as it was misunderstood and largely practiced). Sadly, it misses that Agile was more like "Let's finally admit that iterating haphazardly and unevenly isn't working and instead make it orderly and intentional." than pretending to have invented any of it.
But, you know, shitting on things gets eyeballs. 🤷
Given the stereotypical difficulty of "product folks" and programmers agreeing on and building shared understanding of what to build, this joke seems clear and straightforward. It works because of course, the customer and the programmer failed to agree on something simple.
I use the language that I think my interlocutor would prefer. If I don't know, then I default to the language of the social context, unless I specifically want to practise a language at that moment. If I want to practise, then I ask first.
*whomst'd've.
Come on.