Quantum Computing

by Dave Snyder

Quantum Computing

So-called AI has received a lot of attention over the past couple years. A different type of computing, quantum computing, has not received as much attention.

As was/is the case with so-called AI, quantum computing has been overhyped recently. Quantum computing requires two things, 1) hardware big enough to handle a specific problem and 2) software capable of handling that problem on the given hardware.

There has been slow progress over the past 50 years on quantum hardware, and progress is likely to continue, albeit slowly. Quantum software has been a different story. Some algorithms were created years ago, but adapting these algorithms to a specific situation is frequently tricky. Creating new algorithms is also tricky. There are stubborn issues that don’t occur with conventional software. To date, quantum computers have never been used to solve real life problems.

This may change at some point, but it will probably take a few years if it happens at all.

What is a Quantum Computer?

A quantum computer is a conventional computer with special quantum hardware. The quantum hardware manipulates one or more qubits; a qubit can store information but does so very differently than conventional computers. Each time you add a qubit, you double the amount of information that can be stored. 3 qubits can store twice as much information as 2 qubits; 4 qubits can store twice as much information as 3 qubits and so on. In addition, the stored information can be manipulated very efficiently. In operation, the conventional computer manipulates the qubits, and under the right conditions the combination of conventional and quantum can produce a result faster than a conventional computer alone.

Two properties of specific quantum hardware need to be considered: 1) the number of qubits, the more qubits the better, and 2) the implementation design. There are different ways to implement the quantum hardware each having advantages and disadvantages. These advantages and disadvantages fall into the following categories:

  1. Speed. The speed of a quantum computer depends on the type of hardware. When using a single qubit, quantum hardware is always slower than conventional hardware, but for certain problems the computation can be spread across multiple qubits (for most problems, this is not possible). In such cases, if the number of qubits is large enough, quantum hardware might be faster.
  2. Decoherence time. The information in quantum hardware only persists for a short period of time, the decoherence time. This means any computation must be completed within that time. For specific hardware we can determine an average decoherence time, but it will sometimes shorter and sometimes longer than average. The average decoherence time varies depending on the type of hardware, in most cases it will be less than one second.
  3. Errors. If a computation does not complete on time (which would be expected some fraction of the time), errors result. There are ways to mitigate this problem, but that requires more expensive hardware. The error rate using quantum hardware is always much greater than the error rate using conventional computers.

Details

A quantum computation involves three steps:

  1. The quantum state needs to be set to an initial value. Any data has to be copied from the conventional hardware to the quantum hardware.
  2. The computation takes place.
  3. The results must be measured, transferred from the quantum hardware to the conventional hardware.

Perhaps surprisingly, steps 1 and 3 are the difficult steps. Step 2 is generally straightforward. Given infinite time, step 1 would always be possible. However, the time available is limited. It must be completed withing the decoherence time. Determining what is and what is not possible given these limitations is currently not completely understood but is being researched.

Step 3 is limited by the Heisenberg Uncertainty Principle. It is not possible to transfer all the information from a quantum state to the conventional hardware without some loss. There are ways to deal with this, but it puts hard limits on what is possible.

The bottom line is many potential applications of quantum computers will prove to be impractical and/or better implemented on conventional hardware. The computation must be sufficiently complex, and there must be practical solutions to steps 1 and 3. Only then do the advantages of a quantum computer exceed the disadvantages.

References

See Book List for AI, Machine Learning and Quantum Computing

Links

Modified October 10, 2025