Hacker News Reader

Comment by a3voices | Hacker News Reader

This is ignoring black swan events. There is no chance that police officers will kill 10 million people tomorrow, but there is a non-zero chance that terrorists will.

View on HN

That's completely silly. Yes tail events are important to consider. But major terrorist attack are already events on part of the tail.

All of the successful attacks around the world in the last 15 years probably could have been far more deadly if the terrorists only goal was death count. That's probably not what they're optimizing for, this is framing everything from the victim's point of view. Terrorists can be awful, but they still have motivations and are somewhat rational actors.

I'd be happy to change my view with evidence otherwise, but terrorists are still somewhat "rational actors" with goals that seem horribly irrational to most people. Killing 10 million people in a terrorist attack probably wouldn't serve their long-term goal (even terrorists worry about PR) and would be an incredibly difficult / expensive undertaking.

Dodges the issue entirely. Police kill people in a non-black swan fashion anyway, because their flaws are systemic. We can actually fix the flaws that the cops have.

Terrorists are working as intended, not killing millions of people here but instead squabbling over dirt in a destabilized wasteland while the Russians explode them.

Well there is a chance terrorists will kill millions of people here and it just hasn't happened yet. So the total chance of being killed by a terrorist, if you had more information, might be higher than being killed by an officer.

Hmm, 10 million people is roughly a third of the population of Canada. AFAIK, nothing in recorded history ever killed that many people in a single day. (To give some perspective, the atomic bombs in WWII "only" killed in the range of a hundred thousand people)

If you look at a major terrorist attack death toll, you'll probably find that it's within one order of magnitude of, say, a major riot death toll.

I don't think that this is true. The worst terrorist attacks look to be about an order of magnitude worse than the worst riots.

Looks like I was wrong: riots were worse. Quick google search:

http://www.sickchirpse.com/top-10-most-devastating-terrorist...

http://listverse.com/2014/02/07/10-deadliest-riots-in-histor...

According to those, worst terrorist attack was 9/11, with ~6k deaths in a single day. Most of the riots in the riot list had larger death tolls (ranging from anywhere between ~7k in a single day, to over 1M over a period of time).

No, there isn't such a chance. The only WMDs with the ability to kill tens of thousands are nukes. Terrorists only have nukes on TV, the kind with big countdown displays on them. It's the difference between a black swan and the fantasy of a black swan. The is no real-world hypothesis of how such a black swan happens.

That's why black powder bombs are the highest technology terrorists actually use, because it is the most effective thing they actually can use.

EDIT Before you hypothesize about "stolen nukes" know that it is impossible to move a nuke undetected: http://www.rense.com/general16/nucla.htm

Terrorists and WMDs is fiction. And sometimes fiction is propaganda.

It is possible terrorists could gain access to nukes. You don't know this for sure.

Yes, anyone could know this to a high certainty. It is incredibly difficult to handle and move nuclear materials undetected. It is in fact impossible for non-state actors to deliver nuclear bombs.

This is why every discussion of "nuclear terrorism" also includes "dirty bombs" because otherwise there would be no such topic as "nuclear terrorism."

It may be extremely unlikely, to the point of being negligible, that terrorists could develop, construct and deliver a nuclear device independently. However it may be possible for them to steal one. They'd find it hard to deliver it to a target, but if the theft went undetected (bear in mind that when the USSR disolved there was no definitive inventory of all their nuclear devices and there are still discrepancies in the records) it's certainly not impossible.

There are other ways to kill tens of thousands of people other than nukes.

Weaponized viruses is one. Other indirect ways such as triggering a large scale war, economic warfare, etc.

Another popular-meme way is knocking out the power grid. Take out enough of it and rapid recovery becomes very difficult, soon disrupting & stopping commerce, water, food, gas, etc distribution which some 99% of the US population critically relies on. That often postulated as prompting mass looting & starvation, which one could reasonably assume terminates 3% of the population.

No, there isn't. It is incredibly difficult to make effective chem/bioweapons. Aum tried sarin gas and they would have killed more people with a $10 bomb.

Wars are caused by over-estimating the terrorist threat.

"Terrorists only have nukes on TV"

You sure about that? There's a reason the US gives Pakistan billions in "aid".

There is a non-zero chance that an asteroid will slam into the Earth next month, ending life as we know it. There is a non-zero chance that the LHC will create a black hole that will swallow us all. Etc.

What on earth would make you say there is no (zero) chance police officers will kill 10 million people tomorrow? There is most certainly a chance. It is likely a very, very small chance, but it is still a non-zero chance.

It's a zero chance unless those police officers are also terrorists. Like I said in another comment, there is a zero chance it would happen from routine police activity.

Now you're just playing semantics. There is still some chance it could happen during routine police activity, if that's the game you want to play.

Well perhaps from quantum mechanics there's also a chance that the moon is made out of cheese, but it's not really worth thinking about. The chance that police will kill 10 million people from routine activity is like that. It's not realistic.

Let's assume it is non-zero, but negligible.

Would you say that chance has increased, decreased or remained the same over the last 30 years?

If we are going to sing that song: if somebody is going to kill 10 million people tomorrow is, necessarily, somebody with nuclear or biological weapons.

As a non-American I would feel a lot safe if you just dismantle yours.

American has used nuclear weapons twice in history. Fearing American nukes is ridiculous and not based on any rational thought.

In my opinion, fearing nukes, in general, is pretty rational.

Fear can't be rational. It's definitional. The fear is justifiable in some sense, of course, as we still allow people to feel whatever they want (if not express it)

A counterpoint is that the US is the only country in the world ever to use nuclear weapons. (I don't think that this actually makes the US more likely to use them again.)

1. Last time I checked, 2 > 0.

2. Actually, 2 > 1 also. If the goal was to demonstrate unsurmountable military power and Hiroshima had already happened, what's the point of Nagasaki?

3. Given the past record of violations to international treaties, it can be argued that if the Russians had not developed the capability to retaliate in kind, N >> 2 by now.

Some are arguing that the former case isn't 0% due to extreme fluke events, but they ignore that there are people in the latter case actively trying, right now, to make that chance 100%.

It also ignores that the probability can change over time. Past frequency is not always a good predictor for future probability.

There is a non zero chance a police officer will kill 10 million people tomorrow.

Well that would make him a terrorist. There is a zero chance it would happen from routine police activity.

If you’re talking about really low probability deaths then you need to consider really unusual cases.

Fires gun at fleeing suspect, bullet hits _ -> deaths. Now, how does _ end up with 10 million deaths?

Well, the US Air Force for example has accidently dropped nuclear warheads on several occasions. They are designed to be really tough, but there is probably a non-zero chance a bullet could end up setting a defective one off.

Of course reality can be far stranger than fiction so I am unlikely to think of the most likely risk.

PS: The odds of terrorists being able to kill 10 million people tomorrow is so low it's really not worth considering.

But it's far, far higher chance that terrorists would kill 10 million tomorrow than police. All it takes is for a terrorist group to get their hands on nukes.

If your comparing something with 1 chance in 10^50 vs. 1 chance in 10^55 then sure the first is more likely. It's still not worth considering.

First, getting a nuke while hard is not enough. Killing 10 million people takes either a really big nuke in just the right spot, or a large number of nukes.

Second, having a nuke is not enough you need to be able to correctly detonate on command which it which is harder than you might think.

Third, you need to transport it to a useful location as there not stored in the middle of city's.

Fourth, you need to decide using them is more useful than threating to use them.

Fifth, if your comparing them to US police then they need to target the US.

So, no it takes far more than picking one up to kill 10 million people.

PS: Simply killing lots of people is not that useful to a terrorist organization. Making nukes less useful than you might think.

the chance that a nation state would kill 10 million more tomorrow is several orders of magnitude more probable.

China, US, India, Pakistan, Russia.....

Reply on news.ycombinator.com