The blog post’s Zen approach sounds like an agnostic stance, which probably isn’t a bad thing in that regard.

consider the question: is the distinction between theory A vs B testable?

if we can not yet verify that a distinction is untestable, it might still be testable, but we might just be missing the insight on how to construct the experimental setup to test a distinction.

if in fact a hypothesis eventually turns out to be testable, there is no guarantee that a single person or single group would have come up with such a test, perhaps devising the test requires careful and deliberate thought by a lot of people separated by location and or generations.

in that case, the prescription to "shut up and make testable hypotheses and then invalidate them experimentally" would be counter-productive advice, eliding the agressive "shut up" might do wonders...

as an aside, perhaps we fail to see the forest for the trees:

consider Youngs double slit experiment, the quantum mechanical version, perhaps with single photons or alternatively electrons, self-interfering with single particle detection on a screen.

according to the path integral formalism the particle is taking all possible paths, and the complex amplitudes are added together with their respective path-length-delays, i.e. "simple interference", except I only see course notes picturing the paths through slit A and B, and I never see the quantum mechanics books picturing the infinite number of topological loops: through slit A -> [back through slit B, through slit A again] x N -> to screen...

are many worlds interpretations truly untestable, or are we refusing to acknowledge their having been measured a long time ago?

what prevents a physicist from claiming all of the paths in the path integral formalism correspond to the system histories of the universe splitting over the (continuous or discrete) paths, and then merging by interference on the same final measurement point on the screen.

one can not explain the interference as long as one refuses to recognize the particle passed both slits.

if something so basic to the theory of quantum mechanics such as the single particle equivalent to Youngs double slit experiment could be recognized as not just proof of splitting histories, but also proof of merging histories, that might be seen as a paradigm shift.

what makes self-interference of a particle exploring all paths so much more palatable than simply claiming all these paths correspond to different histories of the photon, and those same histories merging through the laws of interference?

That would render the acceptance of interference into an acceptance of not just many worlds, but also the possibility of diverging histories interacting (at least in setups where interference takes effect).