Will quantum computing replace classical computers?
That’s a popular question, especially as quantum frenzy continues to flood our IT markets.
Will quantum replace classical – NO. The fact that anyone asks that question points to the industry’s less than adequate explanations of and training around what quantum computers will and will not do.
Which is why I wrote this blog 2 years ago and am updating it today:)
Why Won’t Quantum Computers Replace Classical?
It’s simple. Quantum and classical machines aren’t designed to process the same data and applications. At all.
Classical computers process transactional data very well.
- When you need to buy something, sell something, update your revenue projections or headcount costs, or virtually any other data that resides in a database – you will do it with a classical computer. That also includes office productivity applications and more.
- In addition, classical computers will create and store the data that quantum computers analyze. You see, quantum computers have no ability to use a database.
Quantum computers will augment classical computing for the majority of work they will do for the foreseeable future.
- For now, since quantum computers can’t yet scale to process our enormous real-world datasets, we still need classical systems to do the large scale processing.
- Even when quantum computers do scale to meet the vast sizes of the complex data sets they are designed to process, we’ll still use classical systems for the processing work related to anything using a database, for web browsers, for productivity suites like Office and much much more.
- Quantum computers are intended to solve large complex computations, computations that in many cases cannot be completed today by classical systems due to the size and complexity. Think things like optimizing thousands of logistical plans, folding proteins for a model, predicting weather, optimizing troop movements around the globe, analyzing financial data for everythign from credit scores to stock predictions. That’s where quantum will play.
Think back to when Graphics Processing Units (GPUs) were introduced to augment classical systems to speed up computation associated with synthesized graphics generation (vector computation of the 3D scene, shading, texturizing, etc). GPUs didn’t replace classical computers, they enhanced their abilities for the onslaught of gaming, High Performance Computing simulations and more. Sure, there are GPU-heavy systems, but they each need a Central Processing Unit (CPU) to manage processes and the system itself.
The same is true for quantum computers with their Quantum Processing Units (QPUs.) QPUs will serve to augment or enhance classical computers with CPUs for the foreseeable future. Yes, some small experimental problems and simulations can be run on QPUs today. That’s not the same as asking will quantum replace classical computers. At all.
Don’t succumb to the trick question of “will quantum replace classical thinking.” They will exist together as an “and,” not an “or.”
How Will Quantum Augment Classical?
There are many thoughts and scenarios about “hybrid” systems using quantum to augment classical processing.
Quantum computers simulate real world scenarios in ways that classical computers can’t. Their multi-dimensional processing enables complex analysis of complex problems, thanks to capabilities like superposition and entanglement. These features allow them to investigate what happens in complex situations as different elements change, simulating all related changes in the scenario simultaneously.
Classical computers do well at searching the data itself to find the best answer to the problem. The challenge is that they have to work through volumes and volumes of data in a serial fashion, reviewing large amounts of data that often has no relevance to the problem. This is why we hear about problems that run forever and never find a solution. The more dense the data, the more difficult it is for a classical computer to process.
One option for augmenting classical with quantum is to combine their processing in a workflow. For example, the QPU can help locate the most likely positions in the data to find the best answers. The classical computer can then be directed to search for the very best solution or for a diversity of results in those prime data locations.
Another example of blended processing would be the case when “dense” data sectors are sent to a quantum machine for processing, while “sparse” sectors are processed classically. The results can be combined and iterated til the optimum results are computed.
The results are accelerated performance, more accurate results and a diversity of results to choose from.
Where Will Quantum Replace Classical Computers?
It’s most likely that quantum computers in a pure quantum environment will process new, complex simulations and scenarios that we cannot process at all using classical computers today. They’ll open new doors to new thinking and insights we simply can’t imagine today.
That’s why Machine Learning and Artificial Intelligence are areas that are often mentioned with quantum computing. Both technologies need a quantum computer to process at full scale with optimum accuracy and efficiency. Yes, we can process AI and ML today, with limitations. Quantum computing promises to create a breakout scenario where we can process the fullness of these technologies across massive data sets, leveraging complex prompt, parameters and more.
Complex and vast simulations also fall into the potential for “pure quantum.” Fundamental to their nature is quantum computers’ ability to examine a vast number of possible combinations or states very quickly. The first problems to be routinely solved faster entirely on a quantum computer, rather than on a classical computer, may come from many different fields but will almost certainly all exploit this characteristic of quantum computers.
Examples include:
- Packing the most boxes of different sizes into a shipping container
- Analyzing the different ways a molecule can be folded to optimize a chemical reaction
- Finding the most efficient route for a delivery vehicle
- Selecting an optimal financial portfolio based on desired attributes
These problems come from different industries. They all suffer from the difficulty classical computers have with multi-dimensional computations. Because there are so many possible solutions to examine, classical computers tend to lose performance and in many cases, never return and answer, or return it in a timeframe that is not helpful. Since a quantum computer examines many different possibilities simultaneously, problems that will never be practical on a classical computer will become practical on quantum computers.
The Bottom Line
The question of “will quantum computing replace classical computers” is not relevant. There’s a need for both classical and quantum computers, individually and working together.
Each has its own strengths and use cases today. As we expand our quantum knowledge in the future, we’ll find more ways to use quantum with classical, and on its own.
All the hype about quantum computers often misleads people into believing that they are ready for production and will take over the world. That’s not the case. As with any new technology, they need time and vendor effort to evolve to become production-ready. Similarly, they have a specific arena of useful applications, an arena that includes many unsolvable problems and unknown potentials today.
Classical computers will continue to be the workhorses they are today, processing transactions, serving productivity applications and helping us balance our checkbooks.
