North Korea has proven, if anyone had still any doubts, that sanctions and negotiations alone can’t stop a determined dictator from acquiring nukes. In contrast, the jury’s still out on military action. It has only been tried twice, both times by Israel, in Iraq in 1981 and Syria in 2007. And it’s still too soon to say conclusively that it worked. But at least so far, neither country has nuclear weapons.

Moreover, many of the arguments against military action are fatuous. Take, for instance, the claim that military action is pointless once a country has the know-how to build a bomb, because “You can’t bomb a people’s knowledge out of existence,” as New York Times columnist Roger Cohen said of Iran. That’s true, but it’s completely irrelevant. Knowledge is only one of many components needed to build a bomb. Get rid of the others–like Iran’s heavy-water reactor, its stockpile of enriched uranium, and its centrifuges for enriching more–and no amount of knowledge will suffice to produce nuclear weapons.

Then there’s the argument that military action does nothing but buy time. That’s far from self-evident. Some countries might conclude that the effort of rebuilding their nuclear program only to be bombed again isn’t worth it. But even assuming that’s true, buying time has also been proven to be the most sanctions and negotiations can achieve (except in the rare cases where countries actually agree to give up their nuclear programs.

Thus the relevant question is which course of action buys more time, because the more time you buy, the better the chances of an unexpected development—say, regime change in Iran—that could lead to permanent success. Israel’s bombing of Iraq’s nuclear reactor, for instance, bought just enough time for Iraq to make a critical mistake nobody could have foreseen: the invasion of Kuwait in 1990, which led to the Gulf War and America’s subsequent imposition of an intrusive and effective nuclear inspection regime.

For two reasons, military action probably buys the most time. First, sanctions and negotiations leave much of a country’s nuclear infrastructure in place, whereas military action destroys it. Rebuilding from scratch always takes longer than expanding or improving existing infrastructure, especially if military action is combined with sanctions to impede the rebuilding process. Second, unlike military action, negotiations always require concessions, which can actually facilitate nuclear progress by allowing countries to do openly what they would otherwise have to do secretly. The Iran deal, for instance, allows Tehran to replace its old, slow centrifuges with fast new ones, so that when the deal ends—or earlier, if it follows the North Korean model and cheats—it will be able to enrich the uranium needed for a bomb 20 times faster than it could when the deal began.

There is, of course, one serious reason for avoiding military action: fear of painful retaliation. That certainly played a role in America’s reluctance to bomb North Korea; the latter’s conventional forces are quite sufficient to launch devastating reprisals against both South Korea’s civilian population and the tens of thousands of American troops stationed there. Various estimates put the potential casualties in South Korea at tens or even hundreds of thousands.

Low-cost military action was eminently feasible when Iran’s illicit nuclear program was discovered 15 years ago. Unfortunately, that’s no longer true (which is a damning indictment of three successive Israeli governments). Eleven years ago, when Israel fought a month-long war with Tehran’s Lebanese proxy, Hezbollah fired around 4,000 rockets and killed 163 Israelis. Today, Hezbollah has upward of 150,000 rockets, including many with longer ranges, heavier warheads, and greater accuracy. Moreover, back then, Syria had no interest in joining the war, whereas today, it might have little choice. Significant portions of what are euphemistically called “Syrian government forces” are actually militias (both Syrian and foreign) that answer directly to Tehran.

Thus, preparing a military option on Iran starts with taking steps to make this option less dangerous, and therefore more feasible. Those preparations must start with making serious efforts to push Iran out of Syria, curb Iran’s conventional missile program, and persuade Europe to finally outlaw Hezbollah (rather than only its “military wing,” as if this were somehow distinct from its political wing).

All of the above are things America should do anyway to roll back Iranian influence in the Middle East and thereby restore some measure of regional stability. But the nuclear issue gives these steps added urgency.

Most likely, any military action will end up being Israeli rather than American. America has never taken military action to stop any country’s nuclear program, and decision-makers have often gone to extraordinary lengths to avoid doing so. Even though the sanctions/negotiations track has failed repeatedly, Washington is unlikely to scrap this decades-old, bipartisan policy in Iran’s case. A nuclear Iran simply isn’t the existential threat to America that it is to Israel.

But America must begin working now to make Israeli military action feasible at a reasonable cost. For as the North Korean failure shows, only military action is likely to stop Tehran from following in Pyongyang’s footsteps.