Aren't rules of thumb in his case the outcome of inductive reasoning based on many observations? Sounds pretty empirical to me. It is a fundamental strength of science that such formulations accept falsifiability - so the fact that his hypotheses/theories/laws don't hold in all cases just means they need to be refined for those cases.

It also common scientific practice to accept hypotheses within bounds of probability.

The wildest conspiracy theories you've heard are also "the outcome of inductive reasoning based on many observations". Empirical testing is supposed to be statistical, not subject to human bias.

Well no, because they are neither falsifiable nor independently verifiable. Taleb proposes hypotheses based on openly available information and he provides enough information to replicate his findings.