Another way to think about it is, all TSS is equal in the model, but not all TSS is equal in its effect.
Imagine you’re looking at two training weeks, each with 500 TSS. One sample has five rides of toodling around at zone 2 until you hit 100 TSS, the other sample week has three days like that, but two of the days are 100 TSS worth of extremely hard intervals.
Same total TSS but would you expect identical impact on your fitness? I don’t think I would.
I think of it like, CTL is not actually “fitness” in the sense of adaptation. The latter is not directly observable. Rather CTL is your fitness, modeled. And the model has strengths (simple, reasonable correlation with performance, more descriptive than just looking at time or miles) but it also has weaknesses, and this is one.
My 2 cents is the SUF plans play mostly with insensitive and frequency rather than duration. So if you are going add more duration / extra time, definitely introduce it carefully. It’s easier to push yourself over the edge when you have that many workouts that are hard, even if they’re short.