[mikeash]

The core material is the important part if you are, say, Iran or North Korea. However, if you are in a cave with a box of scraps and no Tony Stark, there's a lot more to it.

Exactly how PALs work is not really well known. A lot of well-informed speculation can be found here:

https://www.cs.columbia.edu/~smb/nsam-160/pal.html

The most plausible mechanism for how they work is lots of crypto, plus physical-level crypto-like mechanisms wherein detonation requires setting off the various explosives in the bomb at precise and different times, and those times are not stored in the bomb itself, but rather extracted from the PAL code.

It's possible to build simple nuclear weapons that don't require a lot of complicated steps to explode. But they're tremendously inefficient, and nobody bothers with them today. A stolen nuke will be an implosion design requiring a virtuoso performance of nanosecond-level timing to detonate properly, and those timings will not be easy to reverse engineer. Build a new weapon using the fissile material from the original bomb would likewise be a massive undertaking (only "easy" if you're Iran etc., not al Qaeda), and really would no longer qualify as a "stolen" weapon as proposed.

[nl]

You think the PAL on one of the missing soviet "suitcase" nuclear devices[1] is that sophisticated?

They are booby-trapped against misuse though, so that's ok.

One such cache, identified by Vasili Mitrokhin, exploded when Swiss authorities tried to remove it from a wooded area near Bern.

I don't think civilians know enough about the current state of Soviet-era nuclear devices to be able to assess the risk very well.

Personally I'd be much more concerned about Al-Qaeda linked terrorists getting access to a Pakistani bomb. The Taliban successfully launched a raid on one of Pakistan's main naval bases[2], and NATO also admitted its concern over the Taliban's ability to target the Pakistan's nuclear installations.

Given that the Pakistan Navy noted Al-Qaeda members within its ranks it isn't difficult to imagine a scenario where they get access to the disarm-codes for any PAL-like on the bombs.

I'm not sure if this is an argument for or against domestic surveillance by the NSA, though!

[1] http://en.wikipedia.org/wiki/Suitcase_nuke

[2] http://en.wikipedia.org/wiki/PNS_Mehran_attack

[mikeash]

Are you saying this cache near Bern blew up while containing nuclear weapons? Why haven't I heard of this before? Even with a non-nuclear explosion, such an event would make international news for a month.

[nl]

Allegedly (ie, according to Wikipedia) the details are on pages 475 & 476 of a book called "The KGB in Europe" by former KGB officer Vasili Mitrokhin.

I don't know if that device was nuclear or not - I suspect not.

Edit: some details about other caches:

A former Soviet spy testified at a congressional hearing in Los Angeles on Monday that Russian intelligence operatives placed weapons and communications caches--perhaps even small nuclear devices--in California and other states as part of a plan to destabilize the United States through sabotage

http://articles.latimes.com/2000/jan/25/local/me-57346

[jlgreco]

> The core material is the important part if you are, say, Iran or North Korea. However, if you are in a cave with a box of scraps and no Tony Stark, there's a lot more to it.

Similarly, I suspect that PALs have traditionally been designed with the first in mind. They prevent a proper detonation as the bomb as it currently exists. They seem less concerned with someone using the bomb to cobble together a (relatively speaking) inefficient bomb, or even just a dirty bomb.

If you are satisfied with merely dozens to hundreds dead, and a massive clusterfuck of a disproportionate response, you may have to do nothing more than merely trigger the PAL's anti-tampering mechanism in a populated area. Worse case scenario there is that it doesn't work that way, then you just throw it into a truck with some fertilizer and blow it up anyway.

Nations are not interested in using a bomb like that, they would want the real deal in a form factor that is practical for military use. PALs seem designed to prevent that.

[mikeash]

The possibility of taking an existing bomb and using it to build a new nuclear bomb without being an actual military is absurd. You'll not get an inefficient bomb, you'll get a dud.

Dirty bombs are a different threat altogether. One that's almost entirely psychological. No doubt effective at that, but a completely different scenario than the one being discussed.

[jlgreco]

It is fairly trivial to do with uranium, but with plutonium you are correct; you would never get anything beyond a fizzle (if you were lucky).

If the discussion is "what if terrorists steal a bomb", then the discussion is implicitly about dirty bombs. No other discussion really makes sense.