From what I understand, speculative execution is “worth it” in most non secure contexts. Furthermore, it seems like one could hypothetically implement speculati

From what I understand, speculative execution is “worth it” in most non secure contexts. Furthermore, it seems like one could hypothetically implement speculative execution “correctly”, where speculation is still gated by the constraints of non speculated executions. Could this be a problem that proof software could solve? I still have hope.