So you want to invent a nuclear weapon

>"How to do this? If you shoot two bricks at each other, they’ll likely just bounce. You need to shape the pieces so that they will mesh together. You decide to form one hollow cylinder and one plug that fits perfectly inside the cylinder. You put both of these inside a tube, with an explosive behind the cylinder. When that explosive goes off, the cylinder shoots down the tube and engulfs the plug, establishing a critical mass for fissile reactions to grow exponentially and—boom."

One of the concerns of this kind of nuclear weapon (gun-type) is that, if you don't get the cylinder into the tube fast enough, only part of the mass will undergo fission, resulting in a weak, dirty explosion. The solution the Manhattan project came up with was to basically build something like a single-use howitzer that could be dropped from a plane.

You can also build a hollow sphere of fissile material that isn't supercritical until it is compressed. Then you pack explosives around it that will compress the sphere into a supercritical mass. You need to shape the charge carefully so it compresses evenly and doesn't squirt out somewhere. You also need a neutron source that will get the reaction rolling at the critical moment. You can't just rely on a neutron showing up from the cosmos at just the right moment if your supercritical mass is only assembled for a tiny fraction of a second! Amazingly, this was all solved during WWII, and one of the two bombs dropped in Japan was an implosion bomb.

>"We are very lucky this is so hard."

It's hard, but getting easier all the time. It was hard for the world's most brilliant scientists backed by a state with massive resources to pull off eighty years ago. Now it's merely challenging for nobodies working for third-rate states like North Korea. In a few more decades, it may become possible for non-state actors to build a fission bomb. That's when things will become truly worrisome.