1. Difficult to build. It's proving incredibly hard to achieve any decent range for electrical road vehicles. This multiplies manifold when trying to cram components into a much smaller space and make the whole thing fly.
2. Difficult to control. Gusts of wind, predatory birds, reliable bandwidth etc.
3. Trivially easy to defend against. Think mosquito nets, burqas and cheap radio jammers.
They're building these "bug" micro-drones now. They're doable, they're being done, just a few years to get amazing capabilities implemented.
2. See the video.
3. You might be able to deny access to buildings with mosquito nets but you're not going to stop close range surveillance, painting targets and other uses. Mosquito nets themselves would be relatively trivial to dismantle with a swarm of these (you'd only need to land one with explosives on the net).
Oh? You mean the US government that was the first to create nuclear weapons and land men on the moon?
I say this as someone who's basically libertarian.
A Faraday cage would nullify most of the advantages of the tech and can also be made relatively cheaply.
2. I didn't see any good counter-arguments there. Care to detail.
3. Explosions will alert the building's occupants, and you just sacrificed a complex, expensive device for a $0.50 mosquito net.
Mosquitoes have not only overcome those technical obstacles, but thrived as a species. They even manage to kill millions of people a year, inadvertently. Obviously you can't sweep everything difficult about utilizing these bug-drones under the rug as just questions of implementation, but it's pretty clear that they're not inherently unfeasible.
P.S. Looks like GiraffeNecktie beat me to this, in his comment: https://news.ycombinator.com/item?id=5246511