Update (3/11): This FP column by Graham Allison is to the point.
The U.S. and the USSR had a few close calls during the Cold War, moments when the possibility of a nuclear exchange came too close for comfort. India and Pakistan had an apparently close call during the Kargil crisis in 1999. The existence of these close calls means that nuclear deterrence is not an airtight guarantee against a nuclear exchange. Nor is 'the nuclear taboo' an absolute guarantee, since an exchange between two nuclear-armed countries might conceivably occur essentially unintentionally, i.e. by accident.
However, while nuclear deterrence did not provide an absolute guarantee during the Cold War and while there were a few close calls, on the whole it worked remarkably well, at least in the narrow, relevant sense of "worked". The Cold War never turned into a hot war between the superpowers, who wreaked havoc on the Third World via proxy wars and caused an enormous amount of human misery and death, but managed to avoid the sort of cataclysmic exchange that, in the worst-case scenario, would have meant the end of anything resembling 'civilized' life on the planet. In other words, the "delicate balance of terror" between the superpowers (to borrow the title of a famous article by Albert Wohlstetter from the late 1950s) turned out to be quite sturdy (safety as "the sturdy child of terror," as Churchill put it, in something of a metaphorical mash-up).
These rather unoriginal reflections may serve as a prelude to the thought that, if Iran should one day acquire a nuclear weapon or the capacity to obtain one in a short time frame, the consequence would not be an existential threat to Israel, contrary to Netanyahu's assertions (n.b. I haven't yet read the full transcript of his speech to Congress). Israel of course has its own (officially unacknowledged) nuclear arsenal, and there is every reason to suppose that nuclear deterrence would operate between Iran and Israel as it operates between India and Pakistan, and as it operated between the U.S. and the USSR during the Cold War. That doesn't mean no possibility whatsoever that an exchange could occur, but it suggests it would be highly unlikely. The Iranian leadership would have to be insane to launch a deliberate nuclear strike on a nuclear-armed state with a powerful conventional military, one closely allied to the most militarily powerful country in the world, and I'm aware of no evidence to suggest that Iran's leadership is insane. (Though doubtless there are people who would rush to furnish some were they to read this post, which they probably won't.)
The 'threat' from an Iranian bomb, insofar as there would be one, would come, or so it seems to me, in the form of an increased boldness on Iran's part to throw its weight around in the region, engage in coercive diplomacy vis-a-vis, e.g., Saudi Arabia or Turkey, and generally become more of a nuisance in the eyes of its adversaries. That's not nothing, of course, but it is not the existential threat to Israel that some people claim would be the result of an Iranian nuclear-weapons capacity.
The latest news from the talks is that Iran has rejected the proposal (or 'demand', whichever it was) for a 10-year freeze on certain nuclear activities. However, the talks will and should continue. Netanyahu's prescription of increased sanctions and an end to the negotiations does not seem like a prescription for anything other than disaster in the long run. As Peter T. pointed out in his guest posts recently published on this blog (see here and here), Iran is, by virtue of its size, location, capabilities, and level of development, not the sort of country that can be sanctioned into submission -- not, at least, without setting the stage for precisely the kind of potentially explosive or catastrophic consequences that everyone should be eager to avoid.
A final note about how we think about security claims, which I'm tacking on because I just read Jarrod Hayes's post at Duck of Minerva. Jarrod points out that a speaker's authority to make security claims may be undermined if the claims come to be seen by the target audience as 'political' (though all security claims are political). Although I agree with this, I think Netanyahu's speech is an instance where one should focus on the objective merits of his claims as much as on their 'authority'. Jarrod writes: "Even though the construction of security is intersubjective, it is
spoken about in objective terms. Where the objectivity of the claim rubs
thin, as in Netanyahu’s case, his ability to speak security is undermined."
But the objectivity of Netanyahu's claim rubs thin not simply because it may be perceived as 'political' but because it lacks 'objective' merit. The fact that the construction of security is intersubjective does not mean that there is not a world 'out there' about which one can make better or worse, more plausible or less plausible, claims. The claim that an Iranian nuclear capacity poses an existential threat to Israel is unconvincing, for reasons suggested above. It is unconvincing because it clashes with what history, logic, and evidence suggest about how the real world works. Contrary to Patrick Jackson's view that the world does not exist independently of the mind (mind-world monism), I believe there is a 'real world', that it exists independently of our minds, and that claims about how the world works can be judged as more or less convincing on the basis of evidence. That does not mean I am a neo-positivist (and actually since I have no research agenda and essentially no standing in the IR 'profession', it doesn't really matter what my meta-theoretical leanings are); what it does mean is that in this case we should not lose sight of whether Netanyahu's claims, irrespective of his authority to make them, accord with what we know about the real world.