Which is more likely: that Google’s benchmarking system is wrong, or that quantum computing somehow takes place across hereto unprovable alternate realities?

I know which one I would pick.