[Cryptography] Quantum and continuous progress.

[James A. Donald]

On 2/3/2017 4:00 AM, Tom Mitchell wrote:
> There seems to be continuous progress toward sufficiently interesting
> quantum computers.
>
> http://www.sussex.ac.uk/broadcast/read/38900

"quantum"  and "the size of a large building"

hmm.

The size of a large building?  Something smells funny.


The proposed quantum computer stores its qbits in ions in vacuum traps. 
Presumably the qubits are transported by laser pulses - which requires 
the very nearly lossless single photon readout of the qubit from one 
ion, and corresponding very nearly lossless absorption by another ion. 
Doing anything more interesting than moving data from one storage 
location to another requires very nearly lossless strongly nonlinear 
coupling between photons, or very nearly lossless strongly nonlinear 
coupling between a photon and an ion.

It is in practice hard to get strongly nonlinear coupling unless you 
have distinctly macroscopic laser beams containing billions of photons.

They breathlessly tell us that this is a practical design that could be 
implemented right away, but I am pretty sure we do not have any 
technology to have strongly nonlinear lossless interactions between 
individual photons, nor has anyone even imagined how to do such a thing.

An actually practical quantum computer must necessarily operate at very 
low energy and very high speed, and thus its components for doing logic 
operations must be very small.  We might for example construct large 
molecules that will only break up if the permitted states contain a 
solution to a hard problem - each large molecule would be a quantum 
computer, each large molecule would find the solution, and we would read 
some information about the solution from one molecule, and some more 
information from the next molecule.

Things the size of a building just do not seem likely to preserve much 
quantum character.

You can send one photon containing quantum state through a very long 
fiber optic cable, or you can keep quantum state in a single ion for a 
very long time, but the problem is that to actually do quantum 
computing, you have to have strongly nonlinear nearly lossless 
interactions.  Which is the hard part, the part where very small sizes, 
very high speeds, and very low energies matter.

A component that can do a quantum operation losslessly, can do a 
classical operation very fast at very low energies.  If we had such 
things, we would already be using them for other purposes.