There is no doubt that "functional CSS" resonates strongly with some people. If that term is new to you, I belive it's come to mean the same thing as "Atomic CSS" as defined by John Polacek here. Harry Nicholls likens it to a function that can only produce one result (although I'd call that a pure function or pure component), but instead of a return value being entirely predictable based on inputs, it is an application of style that only does one thing.
I'm of two minds here. People say how fast they can work this way. Great! They like how predictable the applied styles are. Great! I can understand how a tiny stylesheet that doesn't grow over time is appealing as well.
At the same time, I haven't seen writing about other styling concerns. What happens with big redesigns? Is it about the same, time- and difficulty-wise, or do you spend more time tearing down all those classes? What happens when you need a style that isn't available? Write your own? Or does that ruin the spirit of all this and put you in dangerous territory? How intense can all the class names get? I can think of areas I've styled that have three or more media queries that dramatically re-style an element. Putting all that information in HTML seems like it could get awfully messy. Is consistency harder or easier? I get that "p5" might be a useful way to apply an abstract amount of padding, but you still need to sprinkle it all over your codebase and know when to use it, right?
The closest I've been to being convinced about it was hearing from Adam just how configurable Tailwind is. In any case, I find this all endless fascinating and, if you love it, more power to ya.