A secure computer is essentially in a faraday box, and to deal with
   cooling, would be full of mineral oil.
   I have read about Van Eck phreaking lately, and I don't think I see
   much in academia about it. Sure, there's some successful attempts to
   derive RSA keys from laptops, but apparently keyboards are vulnerable (
   [1]https://lasec.epfl.ch/keyboard/ )? LCD monitors leak their
   contents(not a big deal just shelve this away and forget about it)? The
   Inslaw Affair involved a backdoor in Promis using side channel
   emissions?
   Has anyone attempted to see how much signal leaks from telephone lines?
   Or from ethernet cables? Would an algebraic relationship between
   ciphertext bits make it easier for discover the key? Is an MDS matrix
   necessary for linear diffusion?
   Everything is so cheap nowadays, a software defined radio can be had
   for $20.  [2]https://github.com/martinmarinov/TempestSDR
   More impressive than a cryptographic attack against key fobs would be a
   way to eavesdrop on all the communications on a telephone line. That'd
   cause the immediate adoption of the Clipper Chip!
   The academic community for cryptography is strange, no null results are
   published anywhere. It'd be useful to have a casual wiki for those.
   Even if a novel cryptographic technique that requires more plaintext
   than bruteforce for ChaCha is discovered, it would have the potential
   to be improved upon.

References

   1. https://lasec.epfl.ch/keyboard/
   2. https://github.com/martinmarinov/TempestSDR