"Other frontends can choose whether or not they want to use it."
Not to my understanding. This list is implemented via the nodes, and other front ends are using the nodes Steemit provides, according to the very limited grasp I have of this mechanism, and that is why Steempeak and Busy are also being impacted. That's why MIRA could solve the problem.
"...people can still readily access what they did..."
No they can't. People can't extract data from the blockchain anymore than they can build automobiles from scratch. People are functionally limited to publicly available front ends, just as much as they are commercially available cars.
Either trust is a vector for fraud or trustless mechanisms are extant. This list is not a trustless mechanism. You refer to the extant list as if that's the only possible implementation of this mechanism. I tried to point out some humorous examples of why that isn't reasonable, but you seem to have failed to grasp that fact. What if ISIS sent a few thugs to the home office and politely requested some additions or removals from the list? What if CNBC, Russia, or Israel buys Steemit.com? What if hackers attack that centralized single point of weakness?
The status quo won't persist indefinitely, and even if you trust the listmakers today, who makes that list is going to change, as is the list itself. You may not grasp that your trust of those making that list is a vector for fraud, but no victim of fraud ever did until after the fact. I don't distrust @ned to implement this mechanism without fraudulent intent. I just know that @ned isn't any more permanent than I, criminals will seek weaknesses to exploit, and only decentralization prevents centralized power from being projected.