[Serious] Quantum v. Classical computing.

[highdimensionman]

The problem with Quantum computers is noise and hence error and incoherence between qbits. However quantum computing can work in a space where an optimal choice of fields can be made from a large exponential set of fields. kind of like finding an optimal path along a complex light wave array. In order to have all the workings of that wave setup as to find your optimal path on a classical computer you may have to simulate the whole thing where as real world physics can do such things fast without simulation.
When dealing with error quantum engineers may be able to overlap solving so that many instances of clues are gathered multiple times as for the overlapping solutions to amalgamate into a more optimal single solution error free.
Using this approach to solving you get more benefit the better your error correction is per solution thread. 
These exponential fields are directly relative to how many qubits you have so 4 qbits can work with exponentially more fields than 2Qbits.
It's important to note that IBM and Google have yes 56Qbit ect machines but some of the qubits help more in error correction than actually handle the fields so at the moment it's hard to get any kind of coherence at all.
Classical computers need a less centralised data flow and to make use of less accurate more dynamic analogue calculations. Classical computing could also do with going optical.

In terms of how Qbit processes actually compare to classical computer processes there are 2 main possible Qbit types in terms of the computational complexity zoo.

The first way is the one most are working on where you have arrays of qubits in a wave where by the wave becomes exponentially more able to handle certain complex problems the more qbits you have. sometimes you get exponential speed ups sometimes you get lower order speed up's. It's quicker shall we say than a classical computer for some problems.
Psharpe time is the type of problems this type of computer can possibly solve it's not really that good at mimicking a classical computer.
Think highly optimised production efficiency and management of supply and demand or understanding the full dynamic of something as complex as a caffeine molecule

The other way is to use stacked qbits. The problem's may only be exponential to a power of 4 stacks but that's say 1000(Stacked Qbits so each Qbit represents a number between 1 and 1000 not between 1 and 0)^4(the degree of complexity). Working this way you can accurately speed up many classical operations but with polynomial speed up and no exponential speed ups and many other quantum calc benefits would be limited.
P time and below is the type of problem this type of computer may solve faster and it could possibly be used to speed up much classical computation.
Think doom 2040 where you lead into your own nightmare on the holographic display with your mind or an AI chat bot that functions correctly.