In classic Rationalist fashion, she noticed something she didn't like, hypothesized a cause with no real evidence, and then proceeded to rant about the implications of that unproven hypothesis.
My suspicion, on the other hand, is that because Claude is just reproducing statistical patterns from its training data is simply reflecting the fact that she is referred to as a Nazi coder far more often than she is as some kind of open-source luminary. Unless her github metadata signs everything "Justine T the open-source developer" then that association isn't reflected in the patterns extrapolated from the training data.