Time and time again the terrorists turn out to be people who were already being monitored by conventional intelligence strategies. However, the difficulty is in spotting the difference between people with terrorist sympathies and people who are an imminent threat. Harvesting more data won't help with this at all.

Actually, more data probably hurts - adding more "noise" to hunt through.

I'm genuinely curious, if it's not more data, what do we need to differentiate between a would be terrorist and a run of the mill extremist? Like better systems for prioritizing, filtering people?

The argument wasn't that we should be doing things better in order to make these systems "valuable".

The argument was that these types of things simply cannot be done with the current state of tech, and any attempts by governments to do so should be viewed by the public as both useless wastes of public funds and thinly veiled power grabs-- similar to how we view authoritarian regimes that try to extend their own term limits or institute a national curfew in response to political protests.

ahh thank you! So instead of trying to prevent attacks by intercepting them (which is probably doomed to fail anyway), we should work on the greater issue of diplomacy in general.