How a Team of Geeks Cracked the Spy Trade 187
drunken_boxer777 sends us to The Wall Street Journal for a lengthy article on a small tech company, Palantir Technologies, that is making the CIA, Pentagon, and FBI take notice. The submitter adds, "And yes, their company name is a reference to what you think it is." "One of the latest entrants into the government spy-services marketplace, Palantir Technologies has designed what many intelligence analysts say is the most effective tool to date to investigate terrorist networks. The software's main advance is a user-friendly search tool that can scan multiple data sources at once, something previous search tools couldn't do. That means an analyst who is following a tip about a planned terror attack, for example, can more quickly and easily unearth connections among suspects, money transfers, phone calls and previous attacks around the globe. ... With Palantir's software 'you can actually point to examples where it was pretty clear that lives were saved.'"
Call me dense... (Score:4, Interesting)
But what is the reference?
Re:Gee it's almost impressive..... (Score:2, Interesting)
but as TFA points out, the people they're looking for often do things that should get them caught, like using the same address and phone number when buying the plane tickets in the case of the 9/11 hijackers. The basic idea is to find a better way to process the data they already have, and to give people the ability to process data that will help them, even when they don't necessarily have access to it (ie the use of data classified at a level higher than the searcher has access to).
The problem generally hasn't been (so far anyway) that the data wasn't there, it was just that no one had the time or ability to process the information in a useful manner to make these connections. Supposedly this tool does a much better job of it than previous tools, but even if it does, we probably won't hear much more about it either way.
Re:"Perhaps they should call it "One Ring"? (Score:3, Interesting)
They're also disingenuous astroturfers. (Score:3, Interesting)
They've plastered the Pentagon with banners practically claiming they single-handedly brought down GhostNet [wikipedia.org] when they were at best on the periphery of the rather large collection of organizations responsible for it.
Re:Palin? (Score:3, Interesting)
Re:The Palantir Tool is a Double-Edged Sword (Score:2, Interesting)
No software cannot determine your sexual orientation, nor your hidden adultery. It does not read minds, or hearts. It does not magically know events from the past or the future.
I do not know that you can definitively say that. What the tool can ferret out is going to depend on what data it has, how much data it has and the quality of that data.
Does the tool have access to credit card data? Does the tool have access to hotel reservation data? Car rental data? Flight reservation data? Phone records? Movie rental data? Other data types too numerous to list that do exist and that the government might get access to?
Given enough data I don't think anyone can say with absolute certainty that past events cannot be dug out of the data set. Ditto for sussing out probable very near term events based on past events given the proper data in a timely fashion, time to find connections and then bringing those to someone's attention.
Think things like travel plans, especially those made regularly or semi-regularly. Sure, you could say that someone could figure that out on their own, but that would take time. Throw in software that finds those patterns on its own and it becomes a lot easier to do. The software might even shed light on the cause of the travel.
When Person A receives a phone call from number N or an email from address E prior to the last weekend of the month then there is a 75% probability person A will travel to city C and stay in hotel H. Now add in that person B is tied to phone number N and email address E and has the same 75% probability to travel to city C and stay in hotel H...
Now add in a person to see those patterns and look at them and then drill down to the source data that led to the software making those conclusions and also looking into related data.
Think that person might just figure something out? I think there's a pretty good chance they'll learn something.