Back when I was an undergrad at MIT I was a solder monkey for a group that used a dilution refrigerator with a very similar sort of setup, also trying to use current in niobium loops as qbits. However, IIRC we were only able to reach temperatures in the 100mK range, whereas the article claims this group is using temperatures less than 1mK. So I think they would probably have to be using something on top of the dilution refrigerator to get that low.