I’ve heard this critique of the experience of the Carved from Brindlewood approach to mysteries before. I think it’s right insofar as there are limitations to Brindlewood mysteries that Felipe Real’s improvisational mysteries approach does better with—in particular, there’s a lot of prep work Brindlewood requires that improv mysteries don’t need, and the loss of momentum on a failed Theorize roll in Brindlewood can be difficult to manage.
However, I also think it’s incorrect to characterize the Brindlewood framework as “a mystery that you don’t solve; it’s up to the dice.” In fact, in neither improv mysteries nor in Brindlewood do you, the player, solve a mystery. In both, you make up a plausible solution as you improvisationally tell a story about characters solving a mystery. In both, the characters solve a mystery while the players tell a story about the investigation. The solutions are all in-fiction, and not at all at the meta level.
The difference between these frameworks is that it’s impossible to be wrong in improv mysteries. This has advantages—there’s no sudden loss of momentum if a failed solution isn’t handled properly, since the solution can’t fail. But it’s also got disadvantages—there’s no formal opportunity to throw a major twist or complication into the mystery just as it seemed like a solution was in view.
I think both are good ways to handle mysteries in games but they each have their own quirks. I just also think it’s a mistake to describe Brindlewood mysteries as merely “up to the dice.”