You're confusing innovation in features with innovation in basic architecture.
But the premise of the article doesn't even require innovation in features. It just requires changes that change behavior that sites then depend on and that you have to reverse-engineer.
And reverse-engineering is very time-consuming and slow. If all possible competitors have to reverse-engineer to become viable, that puts in place a huge barrier to competition.
Note that WebKit already behaves this way in various cases: their transitions draft proposal was very vague (as in, what they described could have been figured out in a few afternoons by someone playing with the functionality and their developer docs) and then the editors (Apple employees, note) did nothing to improve it for a few years, forcing everyone else to reverse-engineer WebKit to implement this "standard"...