I’m a fallibilist: I don’t believe we can know anything for certain. The best we can do is base propositions off contingent statements: “If what I see is reliable, then what I see in the mirror is not a brain in a vat.”
A brain in a vat is not a very useful starting axioms, so I have no reason to give it serious consideration. By contrast, while taking the general accuracy of my own senses as axiomatic eventually leads me to conclude they can be fallible (example: hallucinations,) it is nonetheless a way more useful axiom for deriving a base of contingent knowledge.