IBM’s latest quantum computer, Heron, just got a significant upgrade in both hardware and software. The company aims to tackle one of the biggest hurdles in the field: error correction. While we’re still a ways from achieving full error correction, Heron does offer error mitigation techniques that help users deal with the inevitable noise from current quantum systems. Essentially, this means software developers need to carefully manage the errors that come with programming these machines.
Jay Gambetta, IBM’s vice president of Quantum, emphasized how advancements in IBM Quantum hardware and the Qiskit software are pushing the envelope. He said they’re enabling users to create new algorithms by tapping into the strengths of both quantum and classical computing resources as they work toward error-corrected systems, which will be crucial for fully exploring the vast opportunities that lie ahead.
Alongside the Heron announcement, IBM unveiled several new tools in its Qiskit developer kit. These tools include the Qiskit Transpiler Service powered by AI for optimizing quantum circuits, and Qiskit Code Assistant, which helps developers generate quantum code using generative AI models. They’re also introducing Qiskit Serverless, allowing developers to execute quantum-centric operations across both quantum and classical systems, along with the Qiskit Functions Catalog that provides access to services from various tech companies.
Tobias Lindstrom from NPL emphasized that true progress in quantum computing hinges on solving the error correction challenge. He believes that once we can create logical error-correcting qubits, scaling up will become much more feasible. He noted that while investing in quantum computers could be expensive, especially if a machine costs around $1 billion, this wouldn’t deter organizations and governments from investing in such a powerful technology.
For now, it seems quantum computers will mainly end up in the hands of large corporations or governments. Lindstrom identified “quantum-type problems,” like those found in quantum chemistry, as ripe for optimization via quantum computing. He also pointed to the UK Research and Innovation’s quantum test bed program as a crucial step towards practical applications of quantum tech, giving firms a chance to develop accessible quantum solutions.
Lindstrom envisions quantum devices functioning like GPUs or FPGAs in high-performance computing. In this model, quantum processors can act as accelerators for CPUs. Ideally, programmers would write their code in familiar languages, while the compiler would identify which tasks would benefit from quantum processing. He described this scenario as the ultimate goal for user-friendly quantum computing.
Despite the ideal vision, Lindstrom highlighted the need for a specialized group of programmers who fully understand quantum architecture. He likened this to the early days of classical computing when mastery over assembly language was necessary for performance optimization. Current industry work aims to make quantum computing more intuitive, but according to Lindstrom, expertise will still be crucial for a while.
As organizations gear up for a world where quantum computing becomes mainstream, Lindstrom pointed out that there’s a pressing need for skills development, similar to the emergence of GPUs. People in tech need to become aware of quantum capabilities without necessarily understanding the intricate details of the hardware itself. They should grasp the application programming interfaces (APIs) and the types of challenges best suited for quantum solutions.