Right now the people who know this cannot answer you without probably breaking NDAs; unless you're lucky enough to find someone who figured it out on his own (which is probably not a good idea because the information would be half-cooked).

This is pretty accurate. I know some EAP people have been very willing to help out with DX12 stuff around here, but generally speaking I'd assume that a lot of the people who are in the EAP or who got DX12 access through other means are sort of in a grey area when it comes to giving out information now the API is in this weird "available to the public, but not really" state; I feel that this is true for me at least, can't speak for anyone else.

Once it's released and once the people under NDA get a green light for speaking about it more openly you'll see a lot more info appear.

Here are the first public D3D12 samples (PSO cache sample included): https://github.com/Microsoft/DirectX-Graphics-Samples

The MSDN documentation is still not fully updated, but looks they are working on it too.

Now that the flood gates are being opened, I can answer.

The source code provided by the post above pretty much covers it. A couple of notes:

1. You still need to fill all the values from the D3D12_GRAPHICS_PIPELINE_STATE_DESC, not just the blob pointer. The DX API is likely unable to read the blob, only the driver, thus it needs you to still fill this data.
2. The cache is unique to a driver & GPU model.
3. Because of the above, you have to check whether PSO creation out of cache succeeded. It can fail if the user updated his drivers or changed his GPU. You can see that the source code provided by MS is doing exactly this.
4. Most of the time is spent compiling HLSL shaders to D3D bytecode ASM, which is an intermediary representation. You should cache this (which is GPU/driver agnostic) and that has been possible since DX11. PSOs go one step further by allowing you to cache the D3D bytecode -> ISA translation, which is usually fast, but still a nice bonus.

Note also that debug VS release build should also invalided the PSO cache, at least the due the Root Layout which should be different...