Question: How hard is it to make an atomic weapon? Answer: Not very.
Back in the 1960s, when proliferation was first becoming a problem, the government decided to take a few Physics post-docs who were completely ignorant of nuclear technology and assign them to build a nuclear device. They would provide support personnel who would manufacture the device and "conduct experiments", the good doctors had to supply the actual design with only their wits and non-classified literature for reference.
And they did. And they didn't just build an easy bomb. They built an implosion device, because the gun model device wasn't a difficult enough problem to be interesting for them.
Now there is a tendency to say that this shows building a nuclear bomb is easy. I'm not going to go that far. Truth be told they had to do a lot of research and many of their experiments would be potentially expensive and dangerous if conducted for real. (Their experiments were basically relayed to experts who would create "correct" results and report them back.) So there is a fair bit of know how and resources that are necessary. Those three brainiacs are the pointy tip of the intellectual spear as it were, but the support staff is still a non-trivial and non-neglectable part of the system. But still it isn't as hard as you might think.
Via In the Agora.
UPDATE: Some people don't know the difference between an implosion device and a gun device. A gun device has a sphere of fissile material with a hole in it. You shoot a plug into the hole to get a critical mass. It's fairly simple and is the type of bomb dropped on Hiroshima. An implosion device is more complicated. The sphere of fissile material is broken up into wedges and a series of explosives are detonated with very precise timing to blow them together and create the critical mass. This is more complicated and trickier.