When Data Stops Being "Data" in Quantum Systems
When we’re talking about spooky action at a distance, data isn’t something we can talk about in the same old ways (store/retrieve, etc.). Instead it’s something we think of as setting up in a system, then shaped so the machine can work through all the possibilities at once to read out a result after it settles into an answer you can actually use.
In the conversation with Chris Powell at SAIC on the most recent episode of the Shared Everything podcast, we delve into this data question in quantum space, covering not just how we (re)define it how information will travel through and between both quantum and hybrid systems.
The conversation gets tricky on the data front because a quantum gate has no register, no place where bits sit and there’s nothing like a file or a block you can point to. Instead, the problem itself has to be expressed directly into the system, usually as a linear algebra structure (a Hamiltonian matrix that can represent all possible states at once, for example). In other words, what we call data becomes a mathematical mapping into a physical system, not something stored inside it.
And as alluded to above, the challenge isn’t storing data but shaping it into a form the quantum system can accept, then translating that result back into something usable after measurement.
This is why everything around quantum today is hybrid, as Powell explains. CPUs and GPUs still do most of the work, preparing data, organizing it, and handling the output. Quantum processors handle specific parts where exploring many states at once is useful and software layers, (Qiskit, Braket, etc.) are translation layers between these two very different models of computing.
We also talk about how traditional data systems are still essential since they hold the data, move it quickly, and repeatedly reshape it into the mathematical forms needed for quantum processing but really, the deeper dive is on how we think about data itself in a quantum system because it’s really closer to a description of a physical system than a set of stored values. You don’t query it in the usual sense, you’re instead setting up conditions so the answer emerges from the system’s behavior.
And as Powell explains, that makes some problems (especially ones that grow too fast for classical systems) much more tractable. It also means the future of quantum data is not about building a new kind of storage, but about learning how to represent problems in a way these systems can actually work with.
Cool stuff at the bleeding edge on today’s Shared Everything episode.
