Professionals don't always get things right, either, and reactionless thrust doesn't really fit into our best current physical model. That doesn't mean it can't exist - but, given the general usefulness of our current model in predicting how things will behave, there's reasonable cause for extraordinary skepticism in response to claims that one of its "can't happen" conditions is inaccurate.

Maybe the paper's conclusion is accurate! Maybe this physical "can't happen" actually can, and the model needs extending to account for that. It wouldn't be the first time. But it also wouldn't be the first time that a "can't happen" really can't happen, and the result suggesting otherwise is an artifact of the way an experiment was run, rather than an accurate description of a previously unsuspected physical phenomenon. Going by past examples, the latter is much more likely than the former. So there's nothing unreasonable, even for people like myself who aren't knowledgeable enough to evaluate the paper on its own merits, in reserving credulity until the result is shown by other experimenters to be reproducible.