A two-year fight to publish a paper suggests quantum computing has a real replication problem

[RayOfLight]

A team led by Sergey Frolov at the University of Pittsburgh, with collaborators from Minnesota and Grenoble, has published a replication study in Science that re-examined some of the most celebrated claims in topological quantum computing, the approach Microsoft has bet two decades on. Their conclusion, the signals that got reported as smoking gun evidence for Majorana zero modes, the exotic particles that would make topological qubits possible, can be explained by ordinary superconducting effects rather than anything genuinely new.

What makes the story land harder is the publishing timeline itself. The paper took roughly two years to get through peer review, and Frolov says earlier replication submissions were repeatedly rejected on the grounds that they lacked novelty, the same structural bias against null results that has fueled replication crises in psychology, biology and medicine. The team eventually had to combine multiple separate replication efforts into one comprehensive paper just to get it taken seriously.

Frolov's own explanation for how this happened is fairly candid, he describes theoretical predictions biasing researchers toward seeing a pattern they were already looking for, convincing themselves this must be it even when a simpler explanation existed. Superconducting nanoscale devices genuinely can produce a lot of different effects, and selective presentation of the more exciting looking data can obscure alternative explanations without anyone necessarily acting in bad faith.

This isn't a new concern in the specific Majorana field either. A high profile Nature paper from 2021 was retracted after Frolov and a colleague raised concerns about data that hadn't originally been included in the published result, so this latest replication study is really the continuation of a fight that has been quietly running for years rather than a sudden new revelation.

None of this means topological quantum computing is finished as an approach. But it does mean a decade of Majorana headlines should probably get filed under open question rather than confirmed, and it's a genuine case study in how commercialization pressure and a journal system biased against replication work can let excitement outrun actual verification

2584476 Quantum Computing May Be Facing A Replication Crisiswww.newscientist.com