Hacker News Reader

Comment by jessriedel | Hacker News Reader

So the obvious problem with the movie Minority Report was the false dichotomy between (a) imperfectly predicting crimes and then punishing people before they commit them and (b) not predicting crimes. The obvious third choice, which has been in use since time immemorial, is (c) imperfectly predicting crime and then applying normal crime-prevention measures to increase the chance of catching the perpetrator and/or reduce the chance that the crime is committed at all, while simultaneously using less resources. In any sane world, the technology of Minority Report would still be used to reduce murders through direct intervention and the deterrent of more effective enforcement, without taking people off to pre-crime prison.

The author's implication that FAST is an implementation of the obviously-bad idea of punishing people for crimes they haven't yet committed is just one of the many bad and misleading aspects of this article.

View on HN

Indeed. I read FAST as a sort of olive-branch in security theatre. It seems that it, or something like it, will be used to remove porno-scanners and grope-checks from standard security. If 70% of people are passing through security without being hassled, their PR problems nosedive.

And frankly, though it's likely to be error prone and leave airport security still largely ineffective, at least it's move toward trying to intelligently identify risks. Rather that simply sending every single person through ineffective security. And lessening the load on the scanners and gropers can only raise their effectiveness.

I agree that that is their goal, but I still don't like rights violated without a warrant or something similar.

As an aside, I find it funny that your name is roc, since http://en.wikipedia.org/wiki/Receiver_operating_characterist... is a good tool to measure accuracy in systems like this.

Yeah, don't get me wrong, I strongly dislike the laundry list of civil liberties infringements since 9/11(TSA/Patriot Act/etc) and I'd much rather see them gone altogether.

But I'm a realist. And if a less-expensive politically-feasible step back toward sanity can be achieved, even if it doesn't put is back across the line directly into sanity by itself, I'm all for taking that step.

70% accuracy is pretty weak. I can predict whether a given passenger will attempt to commit a terrorist act with greater than 99.99% accuracy. Here is my code:

boolean isTerrorist(Passenger p) { return false; }

But this is one of those cases where you're much more interested in recall, than precision, right? The false positive cost is much lower than the false negative cost.

...will be used to remove porno-scanners and grope-checks from standard security.

Has this ever happened? Has the TSA ever given up on a security screening practice that had become a widespread routine?

If 70% of people are passing through security without being hassled, their PR problems nosedive.

If 70% of passengers walk through the screening point without removing their shoes, without walking through a metal detector, without having their bags x-rayed, etc., and the other 30% are singled out for special treatment, won't that 30% feel extra persecuted? The TSA's PR problems are mainly related to rarer-that-30% outrages like selecting comely young women for pornoscanning, interfering with uncommon medical equipment, etc.

> "Has this ever happened?"

No, but I don't think the current state of things is sustainable. There isn't a demographic that's free of TSA horror stories anymore.

And we all know a new vector will be chosen that will defeat these measures. Thus, without a way out, the TSA will have to come up with an even-more-invasive screening method. And I can't see how that's at all feasible.

Drastically reducing the number of people who have to be subject to "advanced" security measures through technology is a face-saving and politically-feasible way for the TSA to back off. It's, frankly, their only way out.

As to the 30% feeling persecuted: yeah, probably. Particularly if a given population consistently gets selected for 'advanced' security. [1] And hopefully that's an avenue that can be leveraged to take further steps toward sane security that doesn't infringe our liberties.

I'm not suggesting this is acceptable as a solution. Just as a step. Provided it is, indeed, a step away from running every single person through the current gauntlet of invasive TSA security theatre.

Children under 12 no longer have to remove their shoes. It's minor, but it is an example of them giving up a routine security screening practice (or at least modifying to be less intrusive).

I think they're going to drop the removing shoes thing completely soon.

>at least it's move toward trying to intelligently identify risks.

If we were looking to intelligently identify risks we would be spending more money on car crashes, "ordinary" murders, tornadoes, earthquakes and other things that kill so many people that we've grown inured to it.

But people are more interested in a magic bullet to stop us from losing 3,000 people in another fell swoop. Which is really a pretty small number given how many untimely deaths we have every year. (Looking at it from a yearly basis, we lose maybe 300 people a year to domestic terrorism. Not a good thing, to be sure, but we have bigger issues.)

The article DOES address this, you just missed it.

> Given that none of these people would have actually committed a terrorist act yet distinguishing the innocent false positives from the guilty might be a non-trivial, and invasive task.

Invading the privacy of (or refusing travel to) the 99.99% false positives IS punishing the innocent. The article does not succumb to the false dichotomy you describe.

That's why I carefully said "implication". He doesn't come out and say it because it would be wrong; he just tries to slander by association.

Furthermore, the FAST data will not result in the refusal of travel (without further evidence) nor will it result in an invasion of privacy which we do not already allow at random. The author is purposefully confusing logically separate issues in order to confuse the reader. The author's argument, as stated, could be directed at any form of intelligence gathering by TSA whatsoever, an obviously wrong conclusion.

One of the subplots was that the technology basically required total subjugation of the precogs. That doesn't seem all that sane.

OK, I agree that sane people could take the principled stand that it's not right to subjugate 3 pre-cogs to save the several hundred people murdered every year in Washington DC. But first, I don't think most people actually would. And second, there has got to be a way to convince 3 pre-cogs to do this work voluntarily, seeing as that you've got about of order a billion dollars (the economic value of 200 lives in the US) to work with.

> In any sane world...

The problem seems to be that we don't live in a sane world.

> So the obvious problem with the movie Minority Report

... was that the screenwriters weren't smart enough to write consistent time-travel logic, and therefore the plot made no sense whatsoever. Also Tom Cruise.

The short story didn't really have a problem of precrime being imperfect per sé.

Reply on news.ycombinator.com