Researchers proposed a qubit made with superfluid helium that could cut quantum computing error rates by roughly 100 times by shielding quantum information from common forms of electromagnetic noise, announced October 1, 2026. Error correction remains one of the central bottlenecks preventing quantum computers from tackling problems beyond current classical computing capability.
The AI relevance here is indirect but real: several of the computational problems quantum computing is expected to eventually help with, large-scale optimization and certain classes of simulation, overlap with problems AI labs already care about, and a 100-fold error reduction, if it holds up outside the lab, would meaningfully move forward the timeline on quantum-assisted AI research specifically.