Imagine, in the aftermath of 9/11, a task force is created with a sizeable budget for catching terrorists before they strike, and sending them to jail. Inside the FBI, talented people are diverted from other investigations, and as time goes by, other careers.

Of course, in the immediate aftermath of 9/11 the biggest question was, will this happen again, how, where and when? Their investigative arm takes the lead to make sure America is safe. But as months go by, oversight gets a little suspicious... with all this money, why have you arrested basically no one?

The experts explain, terrorism is exceptionally rare, you wouldn't expect a lot of arrests, and that helps for a while. But it's a huge budget line item, and over time politicians upset about child pornography complain about funding for TOR bypasses, and other subdivisions complain about the prohibitive cost of investigating and dismantling next-generation drug cartels like The Silk Road. As it becomes clearer the focus of US counter-terrorism will be overseas, the budget in the FBI is at risk unless they can show at least some results. They arrest the people they can, maybe wrap investigations a little earlier than they'd like, hoping they can get people to flip instead of discovering the higher-ups through less revealing means. Some leaders slip through the cracks, maybe. Maybe not, and people constantly worry about it.

Of course, now that they've demonstrated the capacity for convictions, the appetite is hard to suppress. Of course, the Director understands this is worth investing in, even with poor numbers. It was always going to be low volume, and priorities are set in the White House, for the most part. And most of the Congressional oversight understands this too. But to compete with other major priorities, they spend a little bit of time catching the easiest people, to show progress and get money for the real investigations. The ones that, ideally, no one ever hears about, because the terrorism never happens.

The most cost-efficient way to boost numbers is something they might stumble into gradually. A loser online making some dumb statement about wanting to blow something up, taken literally by some agents calling the bluff, until they can send an FBI agent in costume to give them a brick of Play-Doh at a local mall, then scoop them up for attempted terrorism.

It's nonsense, but it helps the department. And over time, they notice oversight rarely digs into composite statistics. Nobody asks questions about capacity and actual risk assessments of potential terror threats – so over time, as priorities shift, it becomes a larger and larger portion of activities. It becomes clear something is wrong – but to fix it requires a higher budget, to follow more tips, which they can only get with numbers indicating it's a bigger problem. So they scoop up more and more essentially harmless losers.

It turns out a former federal prosecutor places the loser-play-doh story at about 85% of modern terrorism arrests. Prosecutors seem willing, if not eager, to shirk their actual duty to spend time protecting the public, violate normal expectations about priorities, engage in tactics that even many lawyers may describe as entrapment, anything so long as a court doesn't throw out the case. The allegation that this process also has an unconstitutional viewpoint bias (as described in the Rolling Stones piece) is almost the smallest problem, and one that's much easier to fix (and may have been fixed already, my understanding is that this stopped). The systemic problem, of incentives and budgeting and specialization leading to career success matching threat modeling – that's a nightmare to fix, and by far the biggest waste I see in modern federal law enforcement.