21 points ·
In the best case, the detection is at the 2.4 sigma confidence level (less than the usual 3 for reasonable confidence and will below the 5 required for strong confidence), and if previously suspected issues with the instrument are true, the detection could disappear completely.
I do not have a strong confidence in this result. (But at least they didn't publish it in Nature, in which case you would absolutely *know* it's wrong...)