>For example, the San Bernardino killers apparently took out a massive loan and emptied their bank accounts prior the attack.

This happens quite frequently when people are going to make a big purchase or pay off another loan with worse terms. It's not as much of a red flag as you would think.

Sure, but taking out a massive loan in coordination with other signs, like buying guns, would start to paint a better picture.

I believe the word is "context."

That done by somebody that the police should be monitoring closely should raise some flags.

Problem is, the police seems to be spending the resources that should be monitoring those people on monitoring everybody.

But coupled with the stated other red flags, this one is bright red with big red arrows pointing to it.

It's hard to say whether they're really "bad at catching the red flags" or not because we're looking at the data after the fact, knowing exactly how the situation turned out. Everything looks like a serious, obvious warning sign once the data has been pre-filtered and given context by the fact that we KNOW these people were planning a terrorist attack.

The other thing to consider is what flags actually allow the authorities to do unless the flags rise to the level where the police can actually arrest the suspects or follow them around 24/7.

Actually, they need both. They are building a ML model that detects probable terrorists. You need to work on fine-turning the model as well as gather more data to improve classification accuracy. The question isn't which approach to take, but how much accuracy is good enough. 100% security is only possible if we monitor every thought of every person on this planet. Are we willing to give up our freedom and privacy for increasing our terrorist detection model performance by 0.1%? (just a guess)

I've seen that movie before, it's called The Minority Report.

It's interesting that you mention The Minority Report. I feel that one of the key themes of the movie (and the story it's based on by Phillip K. Dick) is the question of when someone actually becomes a criminal. That is, if someone is thinking about committing a crime, should they be arrested? Let's say at the last minute they change their mind. If you had arrested that person on the basis that they were going to commit a crime, you would have been wrong and arrested an innocent person.

The point is that we should really consider the wisdom of a surveillance state that we're increasingly tolerating in the US. It is not illegal to be an extremist until you've broken laws. It is illegal for the government to deprive people of their constitutionally guaranteed rights. Like it or not, the rights that we enjoy have to apply to everyone or they aren't worth anything.

This scenario plays out all the time with the exception that it involves 2 or more people: conspiracy to commit a crime.

>> 100% security is only possible if we monitor every thought of every person on this planet

No it isn't because no agency has enough manpower or resources to follow up on every lead a 'terrorist detection model' generates. Catching bad guys isn't all tech. At some point, a human being has to get involved to investigate if they're actually planning to commit an attack.

There is no such thing as 100% security. Maximum security prisons are an example of this fallacy.

The answer is no, we should not be monitoring everyone all the time in the hope that we catch those few small people. I don't want to live in a world where the government is just big brother. We've already seen lots of abuses of the monitoring, like the government targeting reporters of sensitive national security issues (this happened even with Obama). I see surveillance by normal means (phone location, email) getting so pervasive it really will be almost impossible to have anonymous contact with reporters to become a whistle blower - it's almost that today. Maybe only computer programmers will have enough savvy to even attempt anonymity.

Of course there's a purpose in having a government that looks for terrorists - but it is too easy to turn those capabilities into suppression of dissent. Did you see how in Paris they used their new emergency powers to PREEMPTIVELY put people who had protested against climate change in house arrest. Before they even had their convention. That was horrible.

>"For example, the San Bernardino killers apparently took out a massive loan and emptied their bank accounts prior the attack."

The same exact thing happened in the Paris attacks: one of the killers (already known by french authorities to be radicalized) withdrew three months of salary shortly before the assault.

Saying that encryption is the problem is simply ridiculous considering the fact that even the most basic procedures aren't being taken care of.

French authorities didn't manage to stop the Paris attacks despite them also using open, unencrypted communications.